1 # Ocean Interpreter - Jamison Creek version
3 Ocean is intended to be a compiled language, so this interpreter is
4 not targeted at being the final product. It is, rather, an intermediate
5 stage and fills that role in two distinct ways.
7 Firstly, it exists as a platform to experiment with the early language
8 design. An interpreter is easy to write and easy to get working, so
9 the barrier for entry is lower if I aim to start with an interpreter.
11 Secondly, the plan for the Ocean compiler is to write it in the
12 [Ocean language](http://ocean-lang.org). To achieve this we naturally
13 need some sort of boot-strap process and this interpreter - written in
14 portable C - will fill that role. It will be used to bootstrap the
17 Two features that are not needed to fill either of these roles are
18 performance and completeness. The interpreter only needs to be fast
19 enough to run small test programs and occasionally to run the compiler
20 on itself. It only needs to be complete enough to test aspects of the
21 design which are developed before the compiler is working, and to run
22 the compiler on itself. Any features not used by the compiler when
23 compiling itself are superfluous. They may be included anyway, but
26 Nonetheless, the interpreter should end up being reasonably complete,
27 and any performance bottlenecks which appear and are easily fixed, will
32 This third version of the interpreter exists to test out some initial
33 ideas relating to types. Particularly it adds arrays (indexed from
34 zero) and simple structures. Basic control flow and variable scoping
35 are already fairly well established, as are basic numerical and
38 Some operators that have only recently been added, and so have not
39 generated all that much experience yet are "and then" and "or else" as
40 short-circuit Boolean operators, and the "if ... else" trinary
41 operator which can select between two expressions based on a third
42 (which appears syntactically in the middle).
44 The "func" clause currently only allows a "main" function to be
45 declared. That will be extended when proper function support is added.
47 An element that is present purely to make a usable language, and
48 without any expectation that they will remain, is the "print" statement
49 which performs simple output.
51 The current scalar types are "number", "Boolean", and "string".
52 Boolean will likely stay in its current form, the other two might, but
53 could just as easily be changed.
57 Versions of the interpreter which obviously do not support a complete
58 language will be named after creeks and streams. This one is Jamison
61 Once we have something reasonably resembling a complete language, the
62 names of rivers will be used.
63 Early versions of the compiler will be named after seas. Major
64 releases of the compiler will be named after oceans. Hopefully I will
65 be finished once I get to the Pacific Ocean release.
69 As well as parsing and executing a program, the interpreter can print
70 out the program from the parsed internal structure. This is useful
71 for validating the parsing.
72 So the main requirements of the interpreter are:
74 - Parse the program, possibly with tracing,
75 - Analyse the parsed program to ensure consistency,
77 - Execute the "main" function in the program, if no parsing or
78 consistency errors were found.
80 This is all performed by a single C program extracted with
83 There will be two formats for printing the program: a default and one
84 that uses bracketing. So a `--bracket` command line option is needed
85 for that. Normally the first code section found is used, however an
86 alternate section can be requested so that a file (such as this one)
87 can contain multiple programs. This is effected with the `--section`
90 This code must be compiled with `-fplan9-extensions` so that anonymous
91 structures can be used.
93 ###### File: oceani.mk
95 myCFLAGS := -Wall -g -fplan9-extensions
96 CFLAGS := $(filter-out $(myCFLAGS),$(CFLAGS)) $(myCFLAGS)
97 myLDLIBS:= libparser.o libscanner.o libmdcode.o -licuuc
98 LDLIBS := $(filter-out $(myLDLIBS),$(LDLIBS)) $(myLDLIBS)
100 all :: $(LDLIBS) oceani
101 oceani.c oceani.h : oceani.mdc parsergen
102 ./parsergen -o oceani --LALR --tag Parser oceani.mdc
103 oceani.mk: oceani.mdc md2c
106 oceani: oceani.o $(LDLIBS)
107 $(CC) $(CFLAGS) -o oceani oceani.o $(LDLIBS)
109 ###### Parser: header
111 struct parse_context;
113 struct parse_context {
114 struct token_config config;
123 #define container_of(ptr, type, member) ({ \
124 const typeof( ((type *)0)->member ) *__mptr = (ptr); \
125 (type *)( (char *)__mptr - offsetof(type,member) );})
127 #define config2context(_conf) container_of(_conf, struct parse_context, \
130 ###### Parser: reduce
131 struct parse_context *c = config2context(config);
139 #include <sys/mman.h>
158 static char Usage[] =
159 "Usage: oceani --trace --print --noexec --brackets --section=SectionName prog.ocn\n";
160 static const struct option long_options[] = {
161 {"trace", 0, NULL, 't'},
162 {"print", 0, NULL, 'p'},
163 {"noexec", 0, NULL, 'n'},
164 {"brackets", 0, NULL, 'b'},
165 {"section", 1, NULL, 's'},
168 const char *options = "tpnbs";
170 static void pr_err(char *msg) // NOTEST
172 fprintf(stderr, "%s\n", msg); // NOTEST
175 int main(int argc, char *argv[])
180 struct section *s, *ss;
181 char *section = NULL;
182 struct parse_context context = {
184 .ignored = (1 << TK_mark),
185 .number_chars = ".,_+- ",
190 int doprint=0, dotrace=0, doexec=1, brackets=0;
192 while ((opt = getopt_long(argc, argv, options, long_options, NULL))
195 case 't': dotrace=1; break;
196 case 'p': doprint=1; break;
197 case 'n': doexec=0; break;
198 case 'b': brackets=1; break;
199 case 's': section = optarg; break;
200 default: fprintf(stderr, Usage);
204 if (optind >= argc) {
205 fprintf(stderr, "oceani: no input file given\n");
208 fd = open(argv[optind], O_RDONLY);
210 fprintf(stderr, "oceani: cannot open %s\n", argv[optind]);
213 context.file_name = argv[optind];
214 len = lseek(fd, 0, 2);
215 file = mmap(NULL, len, PROT_READ, MAP_SHARED, fd, 0);
216 s = code_extract(file, file+len, pr_err);
218 fprintf(stderr, "oceani: could not find any code in %s\n",
223 ## context initialization
226 for (ss = s; ss; ss = ss->next) {
227 struct text sec = ss->section;
228 if (sec.len == strlen(section) &&
229 strncmp(sec.txt, section, sec.len) == 0)
233 fprintf(stderr, "oceani: cannot find section %s\n",
239 parse_oceani(ss->code, &context.config, dotrace ? stderr : NULL);
242 fprintf(stderr, "oceani: no main function found.\n");
243 context.parse_error = 1;
245 if (context.prog && doprint) {
248 print_exec(context.prog, 0, brackets);
250 if (context.prog && doexec && !context.parse_error) {
251 if (!analyse_prog(context.prog, &context)) {
252 fprintf(stderr, "oceani: type error in program - not running.\n");
255 interp_prog(&context, context.prog, argc - optind, argv+optind);
257 free_exec(context.prog);
260 struct section *t = s->next;
266 ## free context types
267 ## free context storage
268 exit(context.parse_error ? 1 : 0);
273 The four requirements of parse, analyse, print, interpret apply to
274 each language element individually so that is how most of the code
277 Three of the four are fairly self explanatory. The one that requires
278 a little explanation is the analysis step.
280 The current language design does not require the types of variables to
281 be declared, but they must still have a single type. Different
282 operations impose different requirements on the variables, for example
283 addition requires both arguments to be numeric, and assignment
284 requires the variable on the left to have the same type as the
285 expression on the right.
287 Analysis involves propagating these type requirements around and
288 consequently setting the type of each variable. If any requirements
289 are violated (e.g. a string is compared with a number) or if a
290 variable needs to have two different types, then an error is raised
291 and the program will not run.
293 If the same variable is declared in both branchs of an 'if/else', or
294 in all cases of a 'switch' then the multiple instances may be merged
295 into just one variable if the variable is referenced after the
296 conditional statement. When this happens, the types must naturally be
297 consistent across all the branches. When the variable is not used
298 outside the if, the variables in the different branches are distinct
299 and can be of different types.
301 Undeclared names may only appear in "use" statements and "case" expressions.
302 These names are given a type of "label" and a unique value.
303 This allows them to fill the role of a name in an enumerated type, which
304 is useful for testing the `switch` statement.
306 As we will see, the condition part of a `while` statement can return
307 either a Boolean or some other type. This requires that the expected
308 type that gets passed around comprises a type and a flag to indicate
309 that `Tbool` is also permitted.
311 As there are, as yet, no distinct types that are compatible, there
312 isn't much subtlety in the analysis. When we have distinct number
313 types, this will become more interesting.
317 When analysis discovers an inconsistency it needs to report an error;
318 just refusing to run the code ensures that the error doesn't cascade,
319 but by itself it isn't very useful. A clear understanding of the sort
320 of error message that are useful will help guide the process of
323 At a simplistic level, the only sort of error that type analysis can
324 report is that the type of some construct doesn't match a contextual
325 requirement. For example, in `4 + "hello"` the addition provides a
326 contextual requirement for numbers, but `"hello"` is not a number. In
327 this particular example no further information is needed as the types
328 are obvious from local information. When a variable is involved that
329 isn't the case. It may be helpful to explain why the variable has a
330 particular type, by indicating the location where the type was set,
331 whether by declaration or usage.
333 Using a recursive-descent analysis we can easily detect a problem at
334 multiple locations. In "`hello:= "there"; 4 + hello`" the addition
335 will detect that one argument is not a number and the usage of `hello`
336 will detect that a number was wanted, but not provided. In this
337 (early) version of the language, we will generate error reports at
338 multiple locations, so the use of `hello` will report an error and
339 explain were the value was set, and the addition will report an error
340 and say why numbers are needed. To be able to report locations for
341 errors, each language element will need to record a file location
342 (line and column) and each variable will need to record the language
343 element where its type was set. For now we will assume that each line
344 of an error message indicates one location in the file, and up to 2
345 types. So we provide a `printf`-like function which takes a format, a
346 location (a `struct exec` which has not yet been introduced), and 2
347 types. "`%1`" reports the first type, "`%2`" reports the second. We
348 will need a function to print the location, once we know how that is
349 stored. e As will be explained later, there are sometimes extra rules for
350 type matching and they might affect error messages, we need to pass those
353 As well as type errors, we sometimes need to report problems with
354 tokens, which might be unexpected or might name a type that has not
355 been defined. For these we have `tok_err()` which reports an error
356 with a given token. Each of the error functions sets the flag in the
357 context so indicate that parsing failed.
361 static void fput_loc(struct exec *loc, FILE *f);
363 ###### core functions
365 static void type_err(struct parse_context *c,
366 char *fmt, struct exec *loc,
367 struct type *t1, int rules, struct type *t2)
369 fprintf(stderr, "%s:", c->file_name);
370 fput_loc(loc, stderr);
371 for (; *fmt ; fmt++) {
378 case '%': fputc(*fmt, stderr); break; // NOTEST
379 default: fputc('?', stderr); break; // NOTEST
381 type_print(t1, stderr);
384 type_print(t2, stderr);
393 static void tok_err(struct parse_context *c, char *fmt, struct token *t)
395 fprintf(stderr, "%s:%d:%d: %s: %.*s\n", c->file_name, t->line, t->col, fmt,
396 t->txt.len, t->txt.txt);
400 ## Entities: declared and predeclared.
402 There are various "things" that the language and/or the interpreter
403 needs to know about to parse and execute a program. These include
404 types, variables, values, and executable code. These are all lumped
405 together under the term "entities" (calling them "objects" would be
406 confusing) and introduced here. The following section will present the
407 different specific code elements which comprise or manipulate these
412 Values come in a wide range of types, with more likely to be added.
413 Each type needs to be able to print its own values (for convenience at
414 least) as well as to compare two values, at least for equality and
415 possibly for order. For now, values might need to be duplicated and
416 freed, though eventually such manipulations will be better integrated
419 Rather than requiring every numeric type to support all numeric
420 operations (add, multiple, etc), we allow types to be able to present
421 as one of a few standard types: integer, float, and fraction. The
422 existence of these conversion functions eventually enable types to
423 determine if they are compatible with other types, though such types
424 have not yet been implemented.
426 Named type are stored in a simple linked list. Objects of each type are
427 "values" which are often passed around by value.
434 ## value union fields
442 void (*init)(struct type *type, struct value *val);
443 void (*prepare_type)(struct parse_context *c, struct type *type, int parse_time);
444 void (*print)(struct type *type, struct value *val);
445 void (*print_type)(struct type *type, FILE *f);
446 int (*cmp_order)(struct type *t1, struct type *t2,
447 struct value *v1, struct value *v2);
448 int (*cmp_eq)(struct type *t1, struct type *t2,
449 struct value *v1, struct value *v2);
450 void (*dup)(struct type *type, struct value *vold, struct value *vnew);
451 void (*free)(struct type *type, struct value *val);
452 void (*free_type)(struct type *t);
453 long long (*to_int)(struct value *v);
454 double (*to_float)(struct value *v);
455 int (*to_mpq)(mpq_t *q, struct value *v);
464 struct type *typelist;
468 static struct type *find_type(struct parse_context *c, struct text s)
470 struct type *l = c->typelist;
473 text_cmp(l->name, s) != 0)
478 static struct type *add_type(struct parse_context *c, struct text s,
483 n = calloc(1, sizeof(*n));
486 n->next = c->typelist;
491 static void free_type(struct type *t)
493 /* The type is always a reference to something in the
494 * context, so we don't need to free anything.
498 static void free_value(struct type *type, struct value *v)
504 static void type_print(struct type *type, FILE *f)
507 fputs("*unknown*type*", f); // NOTEST
508 else if (type->name.len)
509 fprintf(f, "%.*s", type->name.len, type->name.txt);
510 else if (type->print_type)
511 type->print_type(type, f);
513 fputs("*invalid*type*", f); // NOTEST
516 static void val_init(struct type *type, struct value *val)
518 if (type && type->init)
519 type->init(type, val);
522 static void dup_value(struct type *type,
523 struct value *vold, struct value *vnew)
525 if (type && type->dup)
526 type->dup(type, vold, vnew);
529 static int value_cmp(struct type *tl, struct type *tr,
530 struct value *left, struct value *right)
532 if (tl && tl->cmp_order)
533 return tl->cmp_order(tl, tr, left, right);
534 if (tl && tl->cmp_eq) // NOTEST
535 return tl->cmp_eq(tl, tr, left, right); // NOTEST
539 static void print_value(struct type *type, struct value *v)
541 if (type && type->print)
542 type->print(type, v);
544 printf("*Unknown*"); // NOTEST
549 static void free_value(struct type *type, struct value *v);
550 static int type_compat(struct type *require, struct type *have, int rules);
551 static void type_print(struct type *type, FILE *f);
552 static void val_init(struct type *type, struct value *v);
553 static void dup_value(struct type *type,
554 struct value *vold, struct value *vnew);
555 static int value_cmp(struct type *tl, struct type *tr,
556 struct value *left, struct value *right);
557 static void print_value(struct type *type, struct value *v);
559 ###### free context types
561 while (context.typelist) {
562 struct type *t = context.typelist;
564 context.typelist = t->next;
570 Type can be specified for local variables, for fields in a structure,
571 for formal parameters to functions, and possibly elsewhere. Different
572 rules may apply in different contexts. As a minimum, a named type may
573 always be used. Currently the type of a formal parameter can be
574 different from types in other contexts, so we have a separate grammar
580 Type -> IDENTIFIER ${
581 $0 = find_type(c, $1.txt);
584 "error: undefined type", &$1);
591 FormalType -> Type ${ $0 = $<1; }$
592 ## formal type grammar
596 Values of the base types can be numbers, which we represent as
597 multi-precision fractions, strings, Booleans and labels. When
598 analysing the program we also need to allow for places where no value
599 is meaningful (type `Tnone`) and where we don't know what type to
600 expect yet (type is `NULL`).
602 Values are never shared, they are always copied when used, and freed
603 when no longer needed.
605 When propagating type information around the program, we need to
606 determine if two types are compatible, where type `NULL` is compatible
607 with anything. There are two special cases with type compatibility,
608 both related to the Conditional Statement which will be described
609 later. In some cases a Boolean can be accepted as well as some other
610 primary type, and in others any type is acceptable except a label (`Vlabel`).
611 A separate function encoding these cases will simplify some code later.
613 ###### type functions
615 int (*compat)(struct type *this, struct type *other);
619 static int type_compat(struct type *require, struct type *have, int rules)
621 if ((rules & Rboolok) && have == Tbool)
623 if ((rules & Rnolabel) && have == Tlabel)
625 if (!require || !have)
629 return require->compat(require, have);
631 return require == have;
636 #include "parse_string.h"
637 #include "parse_number.h"
640 myLDLIBS := libnumber.o libstring.o -lgmp
641 LDLIBS := $(filter-out $(myLDLIBS),$(LDLIBS)) $(myLDLIBS)
643 ###### type union fields
644 enum vtype {Vnone, Vstr, Vnum, Vbool, Vlabel} vtype;
646 ###### value union fields
653 static void _free_value(struct type *type, struct value *v)
657 switch (type->vtype) {
659 case Vstr: free(v->str.txt); break;
660 case Vnum: mpq_clear(v->num); break;
666 ###### value functions
668 static void _val_init(struct type *type, struct value *val)
670 switch(type->vtype) {
671 case Vnone: // NOTEST
674 mpq_init(val->num); break;
676 val->str.txt = malloc(1);
688 static void _dup_value(struct type *type,
689 struct value *vold, struct value *vnew)
691 switch (type->vtype) {
692 case Vnone: // NOTEST
695 vnew->label = vold->label;
698 vnew->bool = vold->bool;
702 mpq_set(vnew->num, vold->num);
705 vnew->str.len = vold->str.len;
706 vnew->str.txt = malloc(vnew->str.len);
707 memcpy(vnew->str.txt, vold->str.txt, vnew->str.len);
712 static int _value_cmp(struct type *tl, struct type *tr,
713 struct value *left, struct value *right)
717 return tl - tr; // NOTEST
719 case Vlabel: cmp = left->label == right->label ? 0 : 1; break;
720 case Vnum: cmp = mpq_cmp(left->num, right->num); break;
721 case Vstr: cmp = text_cmp(left->str, right->str); break;
722 case Vbool: cmp = left->bool - right->bool; break;
723 case Vnone: cmp = 0; // NOTEST
728 static void _print_value(struct type *type, struct value *v)
730 switch (type->vtype) {
731 case Vnone: // NOTEST
732 printf("*no-value*"); break; // NOTEST
733 case Vlabel: // NOTEST
734 printf("*label-%p*", v->label); break; // NOTEST
736 printf("%.*s", v->str.len, v->str.txt); break;
738 printf("%s", v->bool ? "True":"False"); break;
743 mpf_set_q(fl, v->num);
744 gmp_printf("%Fg", fl);
751 static void _free_value(struct type *type, struct value *v);
753 static struct type base_prototype = {
755 .print = _print_value,
756 .cmp_order = _value_cmp,
757 .cmp_eq = _value_cmp,
762 static struct type *Tbool, *Tstr, *Tnum, *Tnone, *Tlabel;
765 static struct type *add_base_type(struct parse_context *c, char *n,
766 enum vtype vt, int size)
768 struct text txt = { n, strlen(n) };
771 t = add_type(c, txt, &base_prototype);
774 t->align = size > sizeof(void*) ? sizeof(void*) : size;
775 if (t->size & (t->align - 1))
776 t->size = (t->size | (t->align - 1)) + 1; // NOTEST
780 ###### context initialization
782 Tbool = add_base_type(&context, "Boolean", Vbool, sizeof(char));
783 Tstr = add_base_type(&context, "string", Vstr, sizeof(struct text));
784 Tnum = add_base_type(&context, "number", Vnum, sizeof(mpq_t));
785 Tnone = add_base_type(&context, "none", Vnone, 0);
786 Tlabel = add_base_type(&context, "label", Vlabel, sizeof(void*));
790 Variables are scoped named values. We store the names in a linked list
791 of "bindings" sorted in lexical order, and use sequential search and
798 struct binding *next; // in lexical order
802 This linked list is stored in the parse context so that "reduce"
803 functions can find or add variables, and so the analysis phase can
804 ensure that every variable gets a type.
808 struct binding *varlist; // In lexical order
812 static struct binding *find_binding(struct parse_context *c, struct text s)
814 struct binding **l = &c->varlist;
819 (cmp = text_cmp((*l)->name, s)) < 0)
823 n = calloc(1, sizeof(*n));
830 Each name can be linked to multiple variables defined in different
831 scopes. Each scope starts where the name is declared and continues
832 until the end of the containing code block. Scopes of a given name
833 cannot nest, so a declaration while a name is in-scope is an error.
835 ###### binding fields
836 struct variable *var;
840 struct variable *previous;
842 struct binding *name;
843 struct exec *where_decl;// where name was declared
844 struct exec *where_set; // where type was set
848 While the naming seems strange, we include local constants in the
849 definition of variables. A name declared `var := value` can
850 subsequently be changed, but a name declared `var ::= value` cannot -
853 ###### variable fields
856 Scopes in parallel branches can be partially merged. More
857 specifically, if a given name is declared in both branches of an
858 if/else then its scope is a candidate for merging. Similarly if
859 every branch of an exhaustive switch (e.g. has an "else" clause)
860 declares a given name, then the scopes from the branches are
861 candidates for merging.
863 Note that names declared inside a loop (which is only parallel to
864 itself) are never visible after the loop. Similarly names defined in
865 scopes which are not parallel, such as those started by `for` and
866 `switch`, are never visible after the scope. Only variables defined in
867 both `then` and `else` (including the implicit then after an `if`, and
868 excluding `then` used with `for`) and in all `case`s and `else` of a
869 `switch` or `while` can be visible beyond the `if`/`switch`/`while`.
871 Labels, which are a bit like variables, follow different rules.
872 Labels are not explicitly declared, but if an undeclared name appears
873 in a context where a label is legal, that effectively declares the
874 name as a label. The declaration remains in force (or in scope) at
875 least to the end of the immediately containing block and conditionally
876 in any larger containing block which does not declare the name in some
877 other way. Importantly, the conditional scope extension happens even
878 if the label is only used in one parallel branch of a conditional --
879 when used in one branch it is treated as having been declared in all
882 Merge candidates are tentatively visible beyond the end of the
883 branching statement which creates them. If the name is used, the
884 merge is affirmed and they become a single variable visible at the
885 outer layer. If not - if it is redeclared first - the merge lapses.
887 To track scopes we have an extra stack, implemented as a linked list,
888 which roughly parallels the parse stack and which is used exclusively
889 for scoping. When a new scope is opened, a new frame is pushed and
890 the child-count of the parent frame is incremented. This child-count
891 is used to distinguish between the first of a set of parallel scopes,
892 in which declared variables must not be in scope, and subsequent
893 branches, whether they may already be conditionally scoped.
895 To push a new frame *before* any code in the frame is parsed, we need a
896 grammar reduction. This is most easily achieved with a grammar
897 element which derives the empty string, and creates the new scope when
898 it is recognised. This can be placed, for example, between a keyword
899 like "if" and the code following it.
903 struct scope *parent;
909 struct scope *scope_stack;
912 static void scope_pop(struct parse_context *c)
914 struct scope *s = c->scope_stack;
916 c->scope_stack = s->parent;
921 static void scope_push(struct parse_context *c)
923 struct scope *s = calloc(1, sizeof(*s));
925 c->scope_stack->child_count += 1;
926 s->parent = c->scope_stack;
934 OpenScope -> ${ scope_push(c); }$
935 ClosePara -> ${ var_block_close(c, CloseParallel); }$
937 Each variable records a scope depth and is in one of four states:
939 - "in scope". This is the case between the declaration of the
940 variable and the end of the containing block, and also between
941 the usage with affirms a merge and the end of that block.
943 The scope depth is not greater than the current parse context scope
944 nest depth. When the block of that depth closes, the state will
945 change. To achieve this, all "in scope" variables are linked
946 together as a stack in nesting order.
948 - "pending". The "in scope" block has closed, but other parallel
949 scopes are still being processed. So far, every parallel block at
950 the same level that has closed has declared the name.
952 The scope depth is the depth of the last parallel block that
953 enclosed the declaration, and that has closed.
955 - "conditionally in scope". The "in scope" block and all parallel
956 scopes have closed, and no further mention of the name has been
957 seen. This state includes a secondary nest depth which records the
958 outermost scope seen since the variable became conditionally in
959 scope. If a use of the name is found, the variable becomes "in
960 scope" and that secondary depth becomes the recorded scope depth.
961 If the name is declared as a new variable, the old variable becomes
962 "out of scope" and the recorded scope depth stays unchanged.
964 - "out of scope". The variable is neither in scope nor conditionally
965 in scope. It is permanently out of scope now and can be removed from
966 the "in scope" stack.
968 ###### variable fields
969 int depth, min_depth;
970 enum { OutScope, PendingScope, CondScope, InScope } scope;
971 struct variable *in_scope;
975 struct variable *in_scope;
977 All variables with the same name are linked together using the
978 'previous' link. Those variable that have been affirmatively merged all
979 have a 'merged' pointer that points to one primary variable - the most
980 recently declared instance. When merging variables, we need to also
981 adjust the 'merged' pointer on any other variables that had previously
982 been merged with the one that will no longer be primary.
984 A variable that is no longer the most recent instance of a name may
985 still have "pending" scope, if it might still be merged with most
986 recent instance. These variables don't really belong in the
987 "in_scope" list, but are not immediately removed when a new instance
988 is found. Instead, they are detected and ignored when considering the
989 list of in_scope names.
991 The storage of the value of a variable will be described later. For now
992 we just need to know that when a variable goes out of scope, it might
993 need to be freed. For this we need to be able to find it, so assume that
994 `var_value()` will provide that.
996 ###### variable fields
997 struct variable *merged;
1001 static void variable_merge(struct variable *primary, struct variable *secondary)
1005 if (primary->merged)
1007 primary = primary->merged; // NOTEST
1009 for (v = primary->previous; v; v=v->previous)
1010 if (v == secondary || v == secondary->merged ||
1011 v->merged == secondary ||
1012 (v->merged && v->merged == secondary->merged)) {
1013 v->scope = OutScope;
1014 v->merged = primary;
1018 ###### forward decls
1019 static struct value *var_value(struct parse_context *c, struct variable *v);
1021 ###### free context vars
1023 while (context.varlist) {
1024 struct binding *b = context.varlist;
1025 struct variable *v = b->var;
1026 context.varlist = b->next;
1029 struct variable *t = v;
1032 free_value(t->type, var_value(&context, t));
1034 // This is a global constant
1035 free_exec(t->where_decl);
1040 #### Manipulating Bindings
1042 When a name is conditionally visible, a new declaration discards the
1043 old binding - the condition lapses. Conversely a usage of the name
1044 affirms the visibility and extends it to the end of the containing
1045 block - i.e. the block that contains both the original declaration and
1046 the latest usage. This is determined from `min_depth`. When a
1047 conditionally visible variable gets affirmed like this, it is also
1048 merged with other conditionally visible variables with the same name.
1050 When we parse a variable declaration we either report an error if the
1051 name is currently bound, or create a new variable at the current nest
1052 depth if the name is unbound or bound to a conditionally scoped or
1053 pending-scope variable. If the previous variable was conditionally
1054 scoped, it and its homonyms becomes out-of-scope.
1056 When we parse a variable reference (including non-declarative assignment
1057 "foo = bar") we report an error if the name is not bound or is bound to
1058 a pending-scope variable; update the scope if the name is bound to a
1059 conditionally scoped variable; or just proceed normally if the named
1060 variable is in scope.
1062 When we exit a scope, any variables bound at this level are either
1063 marked out of scope or pending-scoped, depending on whether the scope
1064 was sequential or parallel. Here a "parallel" scope means the "then"
1065 or "else" part of a conditional, or any "case" or "else" branch of a
1066 switch. Other scopes are "sequential".
1068 When exiting a parallel scope we check if there are any variables that
1069 were previously pending and are still visible. If there are, then
1070 there weren't redeclared in the most recent scope, so they cannot be
1071 merged and must become out-of-scope. If it is not the first of
1072 parallel scopes (based on `child_count`), we check that there was a
1073 previous binding that is still pending-scope. If there isn't, the new
1074 variable must now be out-of-scope.
1076 When exiting a sequential scope that immediately enclosed parallel
1077 scopes, we need to resolve any pending-scope variables. If there was
1078 no `else` clause, and we cannot determine that the `switch` was exhaustive,
1079 we need to mark all pending-scope variable as out-of-scope. Otherwise
1080 all pending-scope variables become conditionally scoped.
1083 enum closetype { CloseSequential, CloseParallel, CloseElse };
1085 ###### ast functions
1087 static struct variable *var_decl(struct parse_context *c, struct text s)
1089 struct binding *b = find_binding(c, s);
1090 struct variable *v = b->var;
1092 switch (v ? v->scope : OutScope) {
1094 /* Caller will report the error */
1098 v && v->scope == CondScope;
1100 v->scope = OutScope;
1104 v = calloc(1, sizeof(*v));
1105 v->previous = b->var;
1108 v->min_depth = v->depth = c->scope_depth;
1110 v->in_scope = c->in_scope;
1115 static struct variable *var_ref(struct parse_context *c, struct text s)
1117 struct binding *b = find_binding(c, s);
1118 struct variable *v = b->var;
1119 struct variable *v2;
1121 switch (v ? v->scope : OutScope) {
1124 /* Caller will report the error */
1127 /* All CondScope variables of this name need to be merged
1128 * and become InScope
1130 v->depth = v->min_depth;
1132 for (v2 = v->previous;
1133 v2 && v2->scope == CondScope;
1135 variable_merge(v, v2);
1143 static void var_block_close(struct parse_context *c, enum closetype ct)
1145 /* Close off all variables that are in_scope */
1146 struct variable *v, **vp, *v2;
1149 for (vp = &c->in_scope;
1150 (v = *vp) && v->min_depth > c->scope_depth;
1151 (v->scope == OutScope || v->name->var != v)
1152 ? (*vp = v->in_scope, 0)
1153 : ( vp = &v->in_scope, 0)) {
1154 if (v->name->var != v) {
1155 /* This is still in scope, but we haven't just
1162 case CloseParallel: /* handle PendingScope */
1166 if (c->scope_stack->child_count == 1)
1167 v->scope = PendingScope;
1168 else if (v->previous &&
1169 v->previous->scope == PendingScope)
1170 v->scope = PendingScope;
1171 else if (v->type == Tlabel) // UNTESTED
1172 v->scope = PendingScope; // UNTESTED
1173 else if (v->name->var == v) // UNTESTED
1174 v->scope = OutScope; // UNTESTED
1175 if (ct == CloseElse) {
1176 /* All Pending variables with this name
1177 * are now Conditional */
1179 v2 && v2->scope == PendingScope;
1181 v2->scope = CondScope;
1186 v2 && v2->scope == PendingScope;
1188 if (v2->type != Tlabel)
1189 v2->scope = OutScope;
1191 case OutScope: break; // UNTESTED
1194 case CloseSequential:
1195 if (v->type == Tlabel)
1196 v->scope = PendingScope;
1199 v->scope = OutScope;
1202 /* There was no 'else', so we can only become
1203 * conditional if we know the cases were exhaustive,
1204 * and that doesn't mean anything yet.
1205 * So only labels become conditional..
1208 v2 && v2->scope == PendingScope;
1210 if (v2->type == Tlabel) {
1211 v2->scope = CondScope;
1212 v2->min_depth = c->scope_depth;
1214 v2->scope = OutScope;
1217 case OutScope: break;
1226 The value of a variable is store separately from the variable, on an
1227 analogue of a stack frame. There are (currently) two frames that can be
1228 active. A global frame which currently only stores constants, and a
1229 stacked frame which stores local variables. Each variable knows if it
1230 is global or not, and what its index into the frame is.
1232 Values in the global frame are known immediately they are relevant, so
1233 the frame needs to be reallocated as it grows so it can store those
1234 values. The local frame doesn't get values until the interpreted phase
1235 is started, so there is no need to allocate until the size is known.
1237 ###### variable fields
1241 ###### parse context
1243 short global_size, global_alloc;
1245 void *global, *local;
1247 ###### ast functions
1249 static struct value *var_value(struct parse_context *c, struct variable *v)
1252 if (!c->local || !v->type)
1254 if (v->frame_pos + v->type->size > c->local_size) {
1255 printf("INVALID frame_pos\n"); // NOTEST
1258 return c->local + v->frame_pos;
1260 if (c->global_size > c->global_alloc) {
1261 int old = c->global_alloc;
1262 c->global_alloc = (c->global_size | 1023) + 1024;
1263 c->global = realloc(c->global, c->global_alloc);
1264 memset(c->global + old, 0, c->global_alloc - old);
1266 return c->global + v->frame_pos;
1269 static struct value *global_alloc(struct parse_context *c, struct type *t,
1270 struct variable *v, struct value *init)
1273 struct variable scratch;
1275 if (t->prepare_type)
1276 t->prepare_type(c, t, 1); // NOTEST
1278 if (c->global_size & (t->align - 1))
1279 c->global_size = (c->global_size + t->align) & ~(t->align-1); // UNTESTED
1284 v->frame_pos = c->global_size;
1286 c->global_size += v->type->size;
1287 ret = var_value(c, v);
1289 memcpy(ret, init, t->size);
1295 As global values are found -- struct field initializers, labels etc --
1296 `global_alloc()` is called to record the value in the global frame.
1298 When the program is fully parsed, we need to walk the list of variables
1299 to find any that weren't merged away and that aren't global, and to
1300 calculate the frame size and assign a frame position for each variable.
1301 For this we have `scope_finalize()`.
1303 ###### ast functions
1305 static void scope_finalize(struct parse_context *c)
1309 for (b = c->varlist; b; b = b->next) {
1311 for (v = b->var; v; v = v->previous) {
1312 struct type *t = v->type;
1313 if (v->merged && v->merged != v)
1317 if (c->local_size & (t->align - 1))
1318 c->local_size = (c->local_size + t->align) & ~(t->align-1);
1319 v->frame_pos = c->local_size;
1320 c->local_size += v->type->size;
1323 c->local = calloc(1, c->local_size);
1326 ###### free context storage
1327 free(context.global);
1328 free(context.local);
1332 Executables can be lots of different things. In many cases an
1333 executable is just an operation combined with one or two other
1334 executables. This allows for expressions and lists etc. Other times an
1335 executable is something quite specific like a constant or variable name.
1336 So we define a `struct exec` to be a general executable with a type, and
1337 a `struct binode` which is a subclass of `exec`, forms a node in a
1338 binary tree, and holds an operation. There will be other subclasses,
1339 and to access these we need to be able to `cast` the `exec` into the
1340 various other types. The first field in any `struct exec` is the type
1341 from the `exec_types` enum.
1344 #define cast(structname, pointer) ({ \
1345 const typeof( ((struct structname *)0)->type) *__mptr = &(pointer)->type; \
1346 if (__mptr && *__mptr != X##structname) abort(); \
1347 (struct structname *)( (char *)__mptr);})
1349 #define new(structname) ({ \
1350 struct structname *__ptr = ((struct structname *)calloc(1,sizeof(struct structname))); \
1351 __ptr->type = X##structname; \
1352 __ptr->line = -1; __ptr->column = -1; \
1355 #define new_pos(structname, token) ({ \
1356 struct structname *__ptr = ((struct structname *)calloc(1,sizeof(struct structname))); \
1357 __ptr->type = X##structname; \
1358 __ptr->line = token.line; __ptr->column = token.col; \
1367 enum exec_types type;
1375 struct exec *left, *right;
1378 ###### ast functions
1380 static int __fput_loc(struct exec *loc, FILE *f)
1384 if (loc->line >= 0) {
1385 fprintf(f, "%d:%d: ", loc->line, loc->column);
1388 if (loc->type == Xbinode)
1389 return __fput_loc(cast(binode,loc)->left, f) ||
1390 __fput_loc(cast(binode,loc)->right, f); // NOTEST
1393 static void fput_loc(struct exec *loc, FILE *f)
1395 if (!__fput_loc(loc, f))
1396 fprintf(f, "??:??: "); // NOTEST
1399 Each different type of `exec` node needs a number of functions defined,
1400 a bit like methods. We must be able to free it, print it, analyse it
1401 and execute it. Once we have specific `exec` types we will need to
1402 parse them too. Let's take this a bit more slowly.
1406 The parser generator requires a `free_foo` function for each struct
1407 that stores attributes and they will often be `exec`s and subtypes
1408 there-of. So we need `free_exec` which can handle all the subtypes,
1409 and we need `free_binode`.
1411 ###### ast functions
1413 static void free_binode(struct binode *b)
1418 free_exec(b->right);
1422 ###### core functions
1423 static void free_exec(struct exec *e)
1432 ###### forward decls
1434 static void free_exec(struct exec *e);
1436 ###### free exec cases
1437 case Xbinode: free_binode(cast(binode, e)); break;
1441 Printing an `exec` requires that we know the current indent level for
1442 printing line-oriented components. As will become clear later, we
1443 also want to know what sort of bracketing to use.
1445 ###### ast functions
1447 static void do_indent(int i, char *str)
1454 ###### core functions
1455 static void print_binode(struct binode *b, int indent, int bracket)
1459 ## print binode cases
1463 static void print_exec(struct exec *e, int indent, int bracket)
1469 print_binode(cast(binode, e), indent, bracket); break;
1474 ###### forward decls
1476 static void print_exec(struct exec *e, int indent, int bracket);
1480 As discussed, analysis involves propagating type requirements around the
1481 program and looking for errors.
1483 So `propagate_types` is passed an expected type (being a `struct type`
1484 pointer together with some `val_rules` flags) that the `exec` is
1485 expected to return, and returns the type that it does return, either
1486 of which can be `NULL` signifying "unknown". An `ok` flag is passed
1487 by reference. It is set to `0` when an error is found, and `2` when
1488 any change is made. If it remains unchanged at `1`, then no more
1489 propagation is needed.
1493 enum val_rules {Rnolabel = 1<<0, Rboolok = 1<<1, Rnoconstant = 2<<1};
1497 if (rules & Rnolabel)
1498 fputs(" (labels not permitted)", stderr);
1501 ###### core functions
1503 static struct type *propagate_types(struct exec *prog, struct parse_context *c, int *ok,
1504 struct type *type, int rules);
1505 static struct type *__propagate_types(struct exec *prog, struct parse_context *c, int *ok,
1506 struct type *type, int rules)
1513 switch (prog->type) {
1516 struct binode *b = cast(binode, prog);
1518 ## propagate binode cases
1522 ## propagate exec cases
1527 static struct type *propagate_types(struct exec *prog, struct parse_context *c, int *ok,
1528 struct type *type, int rules)
1530 struct type *ret = __propagate_types(prog, c, ok, type, rules);
1539 Interpreting an `exec` doesn't require anything but the `exec`. State
1540 is stored in variables and each variable will be directly linked from
1541 within the `exec` tree. The exception to this is the `main` function
1542 which needs to look at command line arguments. This function will be
1543 interpreted separately.
1545 Each `exec` can return a value combined with a type in `struct lrval`.
1546 The type may be `Tnone` but must be non-NULL. Some `exec`s will return
1547 the location of a value, which can be updated, in `lval`. Others will
1548 set `lval` to NULL indicating that there is a value of appropriate type
1551 ###### core functions
1555 struct value rval, *lval;
1558 static struct lrval _interp_exec(struct parse_context *c, struct exec *e);
1560 static struct value interp_exec(struct parse_context *c, struct exec *e,
1561 struct type **typeret)
1563 struct lrval ret = _interp_exec(c, e);
1565 if (!ret.type) abort();
1567 *typeret = ret.type;
1569 dup_value(ret.type, ret.lval, &ret.rval);
1573 static struct value *linterp_exec(struct parse_context *c, struct exec *e,
1574 struct type **typeret)
1576 struct lrval ret = _interp_exec(c, e);
1579 *typeret = ret.type;
1581 free_value(ret.type, &ret.rval);
1585 static struct lrval _interp_exec(struct parse_context *c, struct exec *e)
1588 struct value rv = {}, *lrv = NULL;
1589 struct type *rvtype;
1591 rvtype = ret.type = Tnone;
1593 ret.lval = lrv; // UNTESTED
1594 ret.rval = rv; // UNTESTED
1595 return ret; // UNTESTED
1601 struct binode *b = cast(binode, e);
1602 struct value left, right, *lleft;
1603 struct type *ltype, *rtype;
1604 ltype = rtype = Tnone;
1606 ## interp binode cases
1608 free_value(ltype, &left);
1609 free_value(rtype, &right);
1612 ## interp exec cases
1622 Now that we have the shape of the interpreter in place we can add some
1623 complex types and connected them in to the data structures and the
1624 different phases of parse, analyse, print, interpret.
1626 Thus far we have arrays and structs.
1630 Arrays can be declared by giving a size and a type, as `[size]type' so
1631 `freq:[26]number` declares `freq` to be an array of 26 numbers. The
1632 size can be either a literal number, or a named constant. Some day an
1633 arbitrary expression will be supported.
1635 As a formal parameter to a function, the array can be declared with a
1636 new variable as the size: `name:[size::number]string`. The `size`
1637 variable is set to the size of the array and must be a constant. As
1638 `number` is the only supported type, it can be left out:
1639 `name:[size::]string`.
1641 Arrays cannot be assigned. When pointers are introduced we will also
1642 introduce array slices which can refer to part or all of an array -
1643 the assignment syntax will create a slice. For now, an array can only
1644 ever be referenced by the name it is declared with. It is likely that
1645 a "`copy`" primitive will eventually be define which can be used to
1646 make a copy of an array with controllable recursive depth.
1648 For now we have two sorts of array, those with fixed size either because
1649 it is given as a literal number or because it is a struct member (which
1650 cannot have a runtime-changing size), and those with a size that is
1651 determined at runtime - local variables with a const size. The former
1652 have their size calculated at parse time, the latter at run time.
1654 For the latter type, the `size` field of the type is the size of a
1655 pointer, and the array is reallocated every time it comes into scope.
1657 We differentiate struct fields with a const size from local variables
1658 with a const size by whether they are prepared at parse time or not.
1660 ###### type union fields
1663 int unspec; // size is unspecified - vsize must be set.
1666 struct variable *vsize;
1667 struct type *member;
1670 ###### value union fields
1671 void *array; // used if not static_size
1673 ###### value functions
1675 static void array_prepare_type(struct parse_context *c, struct type *type,
1678 struct value *vsize;
1680 if (!type->array.vsize || type->array.static_size)
1683 vsize = var_value(c, type->array.vsize);
1685 mpz_tdiv_q(q, mpq_numref(vsize->num), mpq_denref(vsize->num));
1686 type->array.size = mpz_get_si(q);
1690 type->array.static_size = 1;
1691 type->size = type->array.size * type->array.member->size;
1692 type->align = type->array.member->align;
1696 static void array_init(struct type *type, struct value *val)
1699 void *ptr = val->ptr;
1703 if (!type->array.static_size) {
1704 val->array = calloc(type->array.size,
1705 type->array.member->size);
1708 for (i = 0; i < type->array.size; i++) {
1710 v = (void*)ptr + i * type->array.member->size;
1711 val_init(type->array.member, v);
1715 static void array_free(struct type *type, struct value *val)
1718 void *ptr = val->ptr;
1720 if (!type->array.static_size)
1722 for (i = 0; i < type->array.size; i++) {
1724 v = (void*)ptr + i * type->array.member->size;
1725 free_value(type->array.member, v);
1727 if (!type->array.static_size)
1731 static int array_compat(struct type *require, struct type *have)
1733 if (have->compat != require->compat)
1734 return 0; // UNTESTED
1735 /* Both are arrays, so we can look at details */
1736 if (!type_compat(require->array.member, have->array.member, 0))
1738 if (have->array.unspec && require->array.unspec) {
1739 if (have->array.vsize && require->array.vsize &&
1740 have->array.vsize != require->array.vsize) // UNTESTED
1741 /* sizes might not be the same */
1742 return 0; // UNTESTED
1745 if (have->array.unspec || require->array.unspec)
1746 return 1; // UNTESTED
1747 if (require->array.vsize == NULL && have->array.vsize == NULL)
1748 return require->array.size == have->array.size;
1750 return require->array.vsize == have->array.vsize; // UNTESTED
1753 static void array_print_type(struct type *type, FILE *f)
1756 if (type->array.vsize) {
1757 struct binding *b = type->array.vsize->name;
1758 fprintf(f, "%.*s%s]", b->name.len, b->name.txt,
1759 type->array.unspec ? "::" : "");
1761 fprintf(f, "%d]", type->array.size);
1762 type_print(type->array.member, f);
1765 static struct type array_prototype = {
1767 .prepare_type = array_prepare_type,
1768 .print_type = array_print_type,
1769 .compat = array_compat,
1771 .size = sizeof(void*),
1772 .align = sizeof(void*),
1775 ###### declare terminals
1780 | [ NUMBER ] Type ${ {
1783 struct text noname = { "", 0 };
1786 $0 = t = add_type(c, noname, &array_prototype);
1787 t->array.member = $<4;
1788 t->array.vsize = NULL;
1789 if (number_parse(num, tail, $2.txt) == 0)
1790 tok_err(c, "error: unrecognised number", &$2);
1792 tok_err(c, "error: unsupported number suffix", &$2);
1794 t->array.size = mpz_get_ui(mpq_numref(num));
1795 if (mpz_cmp_ui(mpq_denref(num), 1) != 0) {
1796 tok_err(c, "error: array size must be an integer",
1798 } else if (mpz_cmp_ui(mpq_numref(num), 1UL << 30) >= 0)
1799 tok_err(c, "error: array size is too large",
1803 t->array.static_size = 1;
1804 t->size = t->array.size * t->array.member->size;
1805 t->align = t->array.member->align;
1808 | [ IDENTIFIER ] Type ${ {
1809 struct variable *v = var_ref(c, $2.txt);
1810 struct text noname = { "", 0 };
1813 tok_err(c, "error: name undeclared", &$2);
1814 else if (!v->constant)
1815 tok_err(c, "error: array size must be a constant", &$2);
1817 $0 = add_type(c, noname, &array_prototype);
1818 $0->array.member = $<4;
1820 $0->array.vsize = v;
1825 OptType -> Type ${ $0 = $<1; }$
1828 ###### formal type grammar
1830 | [ IDENTIFIER :: OptType ] Type ${ {
1831 struct variable *v = var_decl(c, $ID.txt);
1832 struct text noname = { "", 0 };
1838 $0 = add_type(c, noname, &array_prototype);
1839 $0->array.member = $<6;
1841 $0->array.unspec = 1;
1842 $0->array.vsize = v;
1848 ###### variable grammar
1850 | Variable [ Expression ] ${ {
1851 struct binode *b = new(binode);
1858 ###### print binode cases
1860 print_exec(b->left, -1, bracket);
1862 print_exec(b->right, -1, bracket);
1866 ###### propagate binode cases
1868 /* left must be an array, right must be a number,
1869 * result is the member type of the array
1871 propagate_types(b->right, c, ok, Tnum, 0);
1872 t = propagate_types(b->left, c, ok, NULL, rules & Rnoconstant);
1873 if (!t || t->compat != array_compat) {
1874 type_err(c, "error: %1 cannot be indexed", prog, t, 0, NULL);
1877 if (!type_compat(type, t->array.member, rules)) {
1878 type_err(c, "error: have %1 but need %2", prog,
1879 t->array.member, rules, type);
1881 return t->array.member;
1885 ###### interp binode cases
1891 lleft = linterp_exec(c, b->left, <ype);
1892 right = interp_exec(c, b->right, &rtype);
1894 mpz_tdiv_q(q, mpq_numref(right.num), mpq_denref(right.num));
1898 if (ltype->array.static_size)
1901 ptr = *(void**)lleft;
1902 rvtype = ltype->array.member;
1903 if (i >= 0 && i < ltype->array.size)
1904 lrv = ptr + i * rvtype->size;
1906 val_init(ltype->array.member, &rv);
1913 A `struct` is a data-type that contains one or more other data-types.
1914 It differs from an array in that each member can be of a different
1915 type, and they are accessed by name rather than by number. Thus you
1916 cannot choose an element by calculation, you need to know what you
1919 The language makes no promises about how a given structure will be
1920 stored in memory - it is free to rearrange fields to suit whatever
1921 criteria seems important.
1923 Structs are declared separately from program code - they cannot be
1924 declared in-line in a variable declaration like arrays can. A struct
1925 is given a name and this name is used to identify the type - the name
1926 is not prefixed by the word `struct` as it would be in C.
1928 Structs are only treated as the same if they have the same name.
1929 Simply having the same fields in the same order is not enough. This
1930 might change once we can create structure initializers from a list of
1933 Each component datum is identified much like a variable is declared,
1934 with a name, one or two colons, and a type. The type cannot be omitted
1935 as there is no opportunity to deduce the type from usage. An initial
1936 value can be given following an equals sign, so
1938 ##### Example: a struct type
1944 would declare a type called "complex" which has two number fields,
1945 each initialised to zero.
1947 Struct will need to be declared separately from the code that uses
1948 them, so we will need to be able to print out the declaration of a
1949 struct when reprinting the whole program. So a `print_type_decl` type
1950 function will be needed.
1952 ###### type union fields
1964 ###### type functions
1965 void (*print_type_decl)(struct type *type, FILE *f);
1967 ###### value functions
1969 static void structure_init(struct type *type, struct value *val)
1973 for (i = 0; i < type->structure.nfields; i++) {
1975 v = (void*) val->ptr + type->structure.fields[i].offset;
1976 if (type->structure.fields[i].init)
1977 dup_value(type->structure.fields[i].type,
1978 type->structure.fields[i].init,
1981 val_init(type->structure.fields[i].type, v);
1985 static void structure_free(struct type *type, struct value *val)
1989 for (i = 0; i < type->structure.nfields; i++) {
1991 v = (void*)val->ptr + type->structure.fields[i].offset;
1992 free_value(type->structure.fields[i].type, v);
1996 static void structure_free_type(struct type *t)
1999 for (i = 0; i < t->structure.nfields; i++)
2000 if (t->structure.fields[i].init) {
2001 free_value(t->structure.fields[i].type,
2002 t->structure.fields[i].init);
2004 free(t->structure.fields);
2007 static struct type structure_prototype = {
2008 .init = structure_init,
2009 .free = structure_free,
2010 .free_type = structure_free_type,
2011 .print_type_decl = structure_print_type,
2025 ###### free exec cases
2027 free_exec(cast(fieldref, e)->left);
2031 ###### declare terminals
2034 ###### variable grammar
2036 | Variable . IDENTIFIER ${ {
2037 struct fieldref *fr = new_pos(fieldref, $2);
2044 ###### print exec cases
2048 struct fieldref *f = cast(fieldref, e);
2049 print_exec(f->left, -1, bracket);
2050 printf(".%.*s", f->name.len, f->name.txt);
2054 ###### ast functions
2055 static int find_struct_index(struct type *type, struct text field)
2058 for (i = 0; i < type->structure.nfields; i++)
2059 if (text_cmp(type->structure.fields[i].name, field) == 0)
2064 ###### propagate exec cases
2068 struct fieldref *f = cast(fieldref, prog);
2069 struct type *st = propagate_types(f->left, c, ok, NULL, 0);
2072 type_err(c, "error: unknown type for field access", f->left, // UNTESTED
2074 else if (st->init != structure_init)
2075 type_err(c, "error: field reference attempted on %1, not a struct",
2076 f->left, st, 0, NULL);
2077 else if (f->index == -2) {
2078 f->index = find_struct_index(st, f->name);
2080 type_err(c, "error: cannot find requested field in %1",
2081 f->left, st, 0, NULL);
2083 if (f->index >= 0) {
2084 struct type *ft = st->structure.fields[f->index].type;
2085 if (!type_compat(type, ft, rules))
2086 type_err(c, "error: have %1 but need %2", prog,
2093 ###### interp exec cases
2096 struct fieldref *f = cast(fieldref, e);
2098 struct value *lleft = linterp_exec(c, f->left, <ype);
2099 lrv = (void*)lleft->ptr + ltype->structure.fields[f->index].offset;
2100 rvtype = ltype->structure.fields[f->index].type;
2106 struct fieldlist *prev;
2110 ###### ast functions
2111 static void free_fieldlist(struct fieldlist *f)
2115 free_fieldlist(f->prev);
2117 free_value(f->f.type, f->f.init); // UNTESTED
2118 free(f->f.init); // UNTESTED
2123 ###### top level grammar
2124 DeclareStruct -> struct IDENTIFIER FieldBlock Newlines ${ {
2126 add_type(c, $2.txt, &structure_prototype);
2128 struct fieldlist *f;
2130 for (f = $3; f; f=f->prev)
2133 t->structure.nfields = cnt;
2134 t->structure.fields = calloc(cnt, sizeof(struct field));
2137 int a = f->f.type->align;
2139 t->structure.fields[cnt] = f->f;
2140 if (t->size & (a-1))
2141 t->size = (t->size | (a-1)) + 1;
2142 t->structure.fields[cnt].offset = t->size;
2143 t->size += ((f->f.type->size - 1) | (a-1)) + 1;
2152 FieldBlock -> { IN OptNL FieldLines OUT OptNL } ${ $0 = $<FL; }$
2153 | { SimpleFieldList } ${ $0 = $<SFL; }$
2154 | IN OptNL FieldLines OUT ${ $0 = $<FL; }$
2155 | SimpleFieldList EOL ${ $0 = $<SFL; }$
2157 FieldLines -> SimpleFieldList Newlines ${ $0 = $<SFL; }$
2158 | FieldLines SimpleFieldList Newlines ${
2163 SimpleFieldList -> Field ${ $0 = $<F; }$
2164 | SimpleFieldList ; Field ${
2168 | SimpleFieldList ; ${
2171 | ERROR ${ tok_err(c, "Syntax error in struct field", &$1); }$
2173 Field -> IDENTIFIER : Type = Expression ${ {
2176 $0 = calloc(1, sizeof(struct fieldlist));
2177 $0->f.name = $1.txt;
2182 propagate_types($<5, c, &ok, $3, 0);
2185 c->parse_error = 1; // UNTESTED
2187 struct value vl = interp_exec(c, $5, NULL);
2188 $0->f.init = global_alloc(c, $0->f.type, NULL, &vl);
2191 | IDENTIFIER : Type ${
2192 $0 = calloc(1, sizeof(struct fieldlist));
2193 $0->f.name = $1.txt;
2195 if ($0->f.type->prepare_type)
2196 $0->f.type->prepare_type(c, $0->f.type, 1);
2199 ###### forward decls
2200 static void structure_print_type(struct type *t, FILE *f);
2202 ###### value functions
2203 static void structure_print_type(struct type *t, FILE *f) // UNTESTED
2207 fprintf(f, "struct %.*s\n", t->name.len, t->name.txt);
2209 for (i = 0; i < t->structure.nfields; i++) {
2210 struct field *fl = t->structure.fields + i;
2211 fprintf(f, " %.*s : ", fl->name.len, fl->name.txt);
2212 type_print(fl->type, f);
2213 if (fl->type->print && fl->init) {
2215 if (fl->type == Tstr)
2216 fprintf(f, "\""); // UNTESTED
2217 print_value(fl->type, fl->init);
2218 if (fl->type == Tstr)
2219 fprintf(f, "\""); // UNTESTED
2225 ###### print type decls
2227 struct type *t; // UNTESTED
2230 while (target != 0) {
2232 for (t = context.typelist; t ; t=t->next)
2233 if (t->print_type_decl) {
2242 t->print_type_decl(t, stdout);
2250 A function is a named chunk of code which can be passed parameters and
2251 can return results. Each function has an implicit type which includes
2252 the set of parameters and the return value. As yet these types cannot
2253 be declared separate from the function itself.
2255 In fact, only one function is currently possible - `main`. `main` is
2256 passed an array of strings together with the size of the array, and
2257 doesn't return anything. The strings are command line arguments.
2259 The parameters can be specified either in parentheses as a list, such as
2261 ##### Example: function 1
2263 func main(av:[ac::number]string)
2266 or as an indented list of one parameter per line
2268 ##### Example: function 2
2271 argv:[argc::number]string
2283 MainFunction -> func main ( OpenScope Args ) Block Newlines ${
2286 $0->left = reorder_bilist($<Ar);
2288 var_block_close(c, CloseSequential);
2289 if (c->scope_stack && !c->parse_error) abort();
2291 | func main IN OpenScope OptNL Args OUT OptNL do Block Newlines ${
2294 $0->left = reorder_bilist($<Ar);
2296 var_block_close(c, CloseSequential);
2297 if (c->scope_stack && !c->parse_error) abort();
2299 | func main NEWLINE OpenScope OptNL do Block Newlines ${
2304 var_block_close(c, CloseSequential);
2305 if (c->scope_stack && !c->parse_error) abort();
2308 Args -> ${ $0 = NULL; }$
2309 | Varlist ${ $0 = $<1; }$
2310 | Varlist ; ${ $0 = $<1; }$
2311 | Varlist NEWLINE ${ $0 = $<1; }$
2313 Varlist -> Varlist ; ArgDecl ${ // UNTESTED
2327 ArgDecl -> IDENTIFIER : FormalType ${ {
2328 struct variable *v = var_decl(c, $1.txt);
2334 ## Executables: the elements of code
2336 Each code element needs to be parsed, printed, analysed,
2337 interpreted, and freed. There are several, so let's just start with
2338 the easy ones and work our way up.
2342 We have already met values as separate objects. When manifest
2343 constants appear in the program text, that must result in an executable
2344 which has a constant value. So the `val` structure embeds a value in
2357 ###### ast functions
2358 struct val *new_val(struct type *T, struct token tk)
2360 struct val *v = new_pos(val, tk);
2371 $0 = new_val(Tbool, $1);
2375 $0 = new_val(Tbool, $1);
2379 $0 = new_val(Tnum, $1);
2382 if (number_parse($0->val.num, tail, $1.txt) == 0)
2383 mpq_init($0->val.num); // UNTESTED
2385 tok_err(c, "error: unsupported number suffix",
2390 $0 = new_val(Tstr, $1);
2393 string_parse(&$1, '\\', &$0->val.str, tail);
2395 tok_err(c, "error: unsupported string suffix",
2400 $0 = new_val(Tstr, $1);
2403 string_parse(&$1, '\\', &$0->val.str, tail);
2405 tok_err(c, "error: unsupported string suffix",
2410 ###### print exec cases
2413 struct val *v = cast(val, e);
2414 if (v->vtype == Tstr)
2416 print_value(v->vtype, &v->val);
2417 if (v->vtype == Tstr)
2422 ###### propagate exec cases
2425 struct val *val = cast(val, prog);
2426 if (!type_compat(type, val->vtype, rules))
2427 type_err(c, "error: expected %1%r found %2",
2428 prog, type, rules, val->vtype);
2432 ###### interp exec cases
2434 rvtype = cast(val, e)->vtype;
2435 dup_value(rvtype, &cast(val, e)->val, &rv);
2438 ###### ast functions
2439 static void free_val(struct val *v)
2442 free_value(v->vtype, &v->val);
2446 ###### free exec cases
2447 case Xval: free_val(cast(val, e)); break;
2449 ###### ast functions
2450 // Move all nodes from 'b' to 'rv', reversing their order.
2451 // In 'b' 'left' is a list, and 'right' is the last node.
2452 // In 'rv', left' is the first node and 'right' is a list.
2453 static struct binode *reorder_bilist(struct binode *b)
2455 struct binode *rv = NULL;
2458 struct exec *t = b->right;
2462 b = cast(binode, b->left);
2472 Just as we used a `val` to wrap a value into an `exec`, we similarly
2473 need a `var` to wrap a `variable` into an exec. While each `val`
2474 contained a copy of the value, each `var` holds a link to the variable
2475 because it really is the same variable no matter where it appears.
2476 When a variable is used, we need to remember to follow the `->merged`
2477 link to find the primary instance.
2485 struct variable *var;
2493 VariableDecl -> IDENTIFIER : ${ {
2494 struct variable *v = var_decl(c, $1.txt);
2495 $0 = new_pos(var, $1);
2500 v = var_ref(c, $1.txt);
2502 type_err(c, "error: variable '%v' redeclared",
2504 type_err(c, "info: this is where '%v' was first declared",
2505 v->where_decl, NULL, 0, NULL);
2508 | IDENTIFIER :: ${ {
2509 struct variable *v = var_decl(c, $1.txt);
2510 $0 = new_pos(var, $1);
2516 v = var_ref(c, $1.txt);
2518 type_err(c, "error: variable '%v' redeclared",
2520 type_err(c, "info: this is where '%v' was first declared",
2521 v->where_decl, NULL, 0, NULL);
2524 | IDENTIFIER : Type ${ {
2525 struct variable *v = var_decl(c, $1.txt);
2526 $0 = new_pos(var, $1);
2533 v = var_ref(c, $1.txt);
2535 type_err(c, "error: variable '%v' redeclared",
2537 type_err(c, "info: this is where '%v' was first declared",
2538 v->where_decl, NULL, 0, NULL);
2541 | IDENTIFIER :: Type ${ {
2542 struct variable *v = var_decl(c, $1.txt);
2543 $0 = new_pos(var, $1);
2551 v = var_ref(c, $1.txt);
2553 type_err(c, "error: variable '%v' redeclared",
2555 type_err(c, "info: this is where '%v' was first declared",
2556 v->where_decl, NULL, 0, NULL);
2561 Variable -> IDENTIFIER ${ {
2562 struct variable *v = var_ref(c, $1.txt);
2563 $0 = new_pos(var, $1);
2565 /* This might be a label - allocate a var just in case */
2566 v = var_decl(c, $1.txt);
2573 cast(var, $0)->var = v;
2577 ###### print exec cases
2580 struct var *v = cast(var, e);
2582 struct binding *b = v->var->name;
2583 printf("%.*s", b->name.len, b->name.txt);
2590 if (loc && loc->type == Xvar) {
2591 struct var *v = cast(var, loc);
2593 struct binding *b = v->var->name;
2594 fprintf(stderr, "%.*s", b->name.len, b->name.txt);
2596 fputs("???", stderr); // NOTEST
2598 fputs("NOTVAR", stderr); // NOTEST
2601 ###### propagate exec cases
2605 struct var *var = cast(var, prog);
2606 struct variable *v = var->var;
2608 type_err(c, "%d:BUG: no variable!!", prog, NULL, 0, NULL); // NOTEST
2609 return Tnone; // NOTEST
2613 if (v->constant && (rules & Rnoconstant)) {
2614 type_err(c, "error: Cannot assign to a constant: %v",
2615 prog, NULL, 0, NULL);
2616 type_err(c, "info: name was defined as a constant here",
2617 v->where_decl, NULL, 0, NULL);
2620 if (v->type == Tnone && v->where_decl == prog)
2621 type_err(c, "error: variable used but not declared: %v",
2622 prog, NULL, 0, NULL);
2623 if (v->type == NULL) {
2624 if (type && *ok != 0) {
2626 v->where_set = prog;
2631 if (!type_compat(type, v->type, rules)) {
2632 type_err(c, "error: expected %1%r but variable '%v' is %2", prog,
2633 type, rules, v->type);
2634 type_err(c, "info: this is where '%v' was set to %1", v->where_set,
2635 v->type, rules, NULL);
2642 ###### interp exec cases
2645 struct var *var = cast(var, e);
2646 struct variable *v = var->var;
2649 v = v->merged; // UNTESTED
2650 lrv = var_value(c, v);
2655 ###### ast functions
2657 static void free_var(struct var *v)
2662 ###### free exec cases
2663 case Xvar: free_var(cast(var, e)); break;
2665 ### Expressions: Conditional
2667 Our first user of the `binode` will be conditional expressions, which
2668 is a bit odd as they actually have three components. That will be
2669 handled by having 2 binodes for each expression. The conditional
2670 expression is the lowest precedence operator which is why we define it
2671 first - to start the precedence list.
2673 Conditional expressions are of the form "value `if` condition `else`
2674 other_value". They associate to the right, so everything to the right
2675 of `else` is part of an else value, while only a higher-precedence to
2676 the left of `if` is the if values. Between `if` and `else` there is no
2677 room for ambiguity, so a full conditional expression is allowed in
2689 Expression -> Expression if Expression else Expression $$ifelse ${ {
2690 struct binode *b1 = new(binode);
2691 struct binode *b2 = new(binode);
2700 ## expression grammar
2702 ###### print binode cases
2705 b2 = cast(binode, b->right);
2706 if (bracket) printf("(");
2707 print_exec(b2->left, -1, bracket);
2709 print_exec(b->left, -1, bracket);
2711 print_exec(b2->right, -1, bracket);
2712 if (bracket) printf(")");
2715 ###### propagate binode cases
2718 /* cond must be Tbool, others must match */
2719 struct binode *b2 = cast(binode, b->right);
2722 propagate_types(b->left, c, ok, Tbool, 0);
2723 t = propagate_types(b2->left, c, ok, type, Rnolabel);
2724 t2 = propagate_types(b2->right, c, ok, type ?: t, Rnolabel);
2728 ###### interp binode cases
2731 struct binode *b2 = cast(binode, b->right);
2732 left = interp_exec(c, b->left, <ype);
2734 rv = interp_exec(c, b2->left, &rvtype); // UNTESTED
2736 rv = interp_exec(c, b2->right, &rvtype);
2740 ### Expressions: Boolean
2742 The next class of expressions to use the `binode` will be Boolean
2743 expressions. "`and then`" and "`or else`" are similar to `and` and `or`
2744 have same corresponding precendence. The difference is that they don't
2745 evaluate the second expression if not necessary.
2754 ###### expr precedence
2759 ###### expression grammar
2760 | Expression or Expression ${ {
2761 struct binode *b = new(binode);
2767 | Expression or else Expression ${ {
2768 struct binode *b = new(binode);
2775 | Expression and Expression ${ {
2776 struct binode *b = new(binode);
2782 | Expression and then Expression ${ {
2783 struct binode *b = new(binode);
2790 | not Expression ${ {
2791 struct binode *b = new(binode);
2797 ###### print binode cases
2799 if (bracket) printf("(");
2800 print_exec(b->left, -1, bracket);
2802 print_exec(b->right, -1, bracket);
2803 if (bracket) printf(")");
2806 if (bracket) printf("(");
2807 print_exec(b->left, -1, bracket);
2808 printf(" and then ");
2809 print_exec(b->right, -1, bracket);
2810 if (bracket) printf(")");
2813 if (bracket) printf("(");
2814 print_exec(b->left, -1, bracket);
2816 print_exec(b->right, -1, bracket);
2817 if (bracket) printf(")");
2820 if (bracket) printf("(");
2821 print_exec(b->left, -1, bracket);
2822 printf(" or else ");
2823 print_exec(b->right, -1, bracket);
2824 if (bracket) printf(")");
2827 if (bracket) printf("(");
2829 print_exec(b->right, -1, bracket);
2830 if (bracket) printf(")");
2833 ###### propagate binode cases
2839 /* both must be Tbool, result is Tbool */
2840 propagate_types(b->left, c, ok, Tbool, 0);
2841 propagate_types(b->right, c, ok, Tbool, 0);
2842 if (type && type != Tbool)
2843 type_err(c, "error: %1 operation found where %2 expected", prog,
2847 ###### interp binode cases
2849 rv = interp_exec(c, b->left, &rvtype);
2850 right = interp_exec(c, b->right, &rtype);
2851 rv.bool = rv.bool && right.bool;
2854 rv = interp_exec(c, b->left, &rvtype);
2856 rv = interp_exec(c, b->right, NULL);
2859 rv = interp_exec(c, b->left, &rvtype);
2860 right = interp_exec(c, b->right, &rtype);
2861 rv.bool = rv.bool || right.bool;
2864 rv = interp_exec(c, b->left, &rvtype);
2866 rv = interp_exec(c, b->right, NULL);
2869 rv = interp_exec(c, b->right, &rvtype);
2873 ### Expressions: Comparison
2875 Of slightly higher precedence that Boolean expressions are Comparisons.
2876 A comparison takes arguments of any comparable type, but the two types
2879 To simplify the parsing we introduce an `eop` which can record an
2880 expression operator, and the `CMPop` non-terminal will match one of them.
2887 ###### ast functions
2888 static void free_eop(struct eop *e)
2902 ###### expr precedence
2903 $LEFT < > <= >= == != CMPop
2905 ###### expression grammar
2906 | Expression CMPop Expression ${ {
2907 struct binode *b = new(binode);
2917 CMPop -> < ${ $0.op = Less; }$
2918 | > ${ $0.op = Gtr; }$
2919 | <= ${ $0.op = LessEq; }$
2920 | >= ${ $0.op = GtrEq; }$
2921 | == ${ $0.op = Eql; }$
2922 | != ${ $0.op = NEql; }$
2924 ###### print binode cases
2932 if (bracket) printf("(");
2933 print_exec(b->left, -1, bracket);
2935 case Less: printf(" < "); break;
2936 case LessEq: printf(" <= "); break;
2937 case Gtr: printf(" > "); break;
2938 case GtrEq: printf(" >= "); break;
2939 case Eql: printf(" == "); break;
2940 case NEql: printf(" != "); break;
2941 default: abort(); // NOTEST
2943 print_exec(b->right, -1, bracket);
2944 if (bracket) printf(")");
2947 ###### propagate binode cases
2954 /* Both must match but not be labels, result is Tbool */
2955 t = propagate_types(b->left, c, ok, NULL, Rnolabel);
2957 propagate_types(b->right, c, ok, t, 0);
2959 t = propagate_types(b->right, c, ok, NULL, Rnolabel); // UNTESTED
2961 t = propagate_types(b->left, c, ok, t, 0); // UNTESTED
2963 if (!type_compat(type, Tbool, 0))
2964 type_err(c, "error: Comparison returns %1 but %2 expected", prog,
2965 Tbool, rules, type);
2968 ###### interp binode cases
2977 left = interp_exec(c, b->left, <ype);
2978 right = interp_exec(c, b->right, &rtype);
2979 cmp = value_cmp(ltype, rtype, &left, &right);
2982 case Less: rv.bool = cmp < 0; break;
2983 case LessEq: rv.bool = cmp <= 0; break;
2984 case Gtr: rv.bool = cmp > 0; break;
2985 case GtrEq: rv.bool = cmp >= 0; break;
2986 case Eql: rv.bool = cmp == 0; break;
2987 case NEql: rv.bool = cmp != 0; break;
2988 default: rv.bool = 0; break; // NOTEST
2993 ### Expressions: The rest
2995 The remaining expressions with the highest precedence are arithmetic,
2996 string concatenation, and string conversion. String concatenation
2997 (`++`) has the same precedence as multiplication and division, but lower
3000 String conversion is a temporary feature until I get a better type
3001 system. `$` is a prefix operator which expects a string and returns
3004 `+` and `-` are both infix and prefix operations (where they are
3005 absolute value and negation). These have different operator names.
3007 We also have a 'Bracket' operator which records where parentheses were
3008 found. This makes it easy to reproduce these when printing. Possibly I
3009 should only insert brackets were needed for precedence.
3019 ###### expr precedence
3025 ###### expression grammar
3026 | Expression Eop Expression ${ {
3027 struct binode *b = new(binode);
3034 | Expression Top Expression ${ {
3035 struct binode *b = new(binode);
3042 | ( Expression ) ${ {
3043 struct binode *b = new_pos(binode, $1);
3048 | Uop Expression ${ {
3049 struct binode *b = new(binode);
3054 | Value ${ $0 = $<1; }$
3055 | Variable ${ $0 = $<1; }$
3058 Eop -> + ${ $0.op = Plus; }$
3059 | - ${ $0.op = Minus; }$
3061 Uop -> + ${ $0.op = Absolute; }$
3062 | - ${ $0.op = Negate; }$
3063 | $ ${ $0.op = StringConv; }$
3065 Top -> * ${ $0.op = Times; }$
3066 | / ${ $0.op = Divide; }$
3067 | % ${ $0.op = Rem; }$
3068 | ++ ${ $0.op = Concat; }$
3070 ###### print binode cases
3077 if (bracket) printf("(");
3078 print_exec(b->left, indent, bracket);
3080 case Plus: fputs(" + ", stdout); break;
3081 case Minus: fputs(" - ", stdout); break;
3082 case Times: fputs(" * ", stdout); break;
3083 case Divide: fputs(" / ", stdout); break;
3084 case Rem: fputs(" % ", stdout); break;
3085 case Concat: fputs(" ++ ", stdout); break;
3086 default: abort(); // NOTEST
3088 print_exec(b->right, indent, bracket);
3089 if (bracket) printf(")");
3094 if (bracket) printf("(");
3096 case Absolute: fputs("+", stdout); break;
3097 case Negate: fputs("-", stdout); break;
3098 case StringConv: fputs("$", stdout); break;
3099 default: abort(); // NOTEST
3101 print_exec(b->right, indent, bracket);
3102 if (bracket) printf(")");
3106 print_exec(b->right, indent, bracket);
3110 ###### propagate binode cases
3116 /* both must be numbers, result is Tnum */
3119 /* as propagate_types ignores a NULL,
3120 * unary ops fit here too */
3121 propagate_types(b->left, c, ok, Tnum, 0);
3122 propagate_types(b->right, c, ok, Tnum, 0);
3123 if (!type_compat(type, Tnum, 0))
3124 type_err(c, "error: Arithmetic returns %1 but %2 expected", prog,
3129 /* both must be Tstr, result is Tstr */
3130 propagate_types(b->left, c, ok, Tstr, 0);
3131 propagate_types(b->right, c, ok, Tstr, 0);
3132 if (!type_compat(type, Tstr, 0))
3133 type_err(c, "error: Concat returns %1 but %2 expected", prog,
3138 /* op must be string, result is number */
3139 propagate_types(b->left, c, ok, Tstr, 0);
3140 if (!type_compat(type, Tnum, 0))
3141 type_err(c, // UNTESTED
3142 "error: Can only convert string to number, not %1",
3143 prog, type, 0, NULL);
3147 return propagate_types(b->right, c, ok, type, 0);
3149 ###### interp binode cases
3152 rv = interp_exec(c, b->left, &rvtype);
3153 right = interp_exec(c, b->right, &rtype);
3154 mpq_add(rv.num, rv.num, right.num);
3157 rv = interp_exec(c, b->left, &rvtype);
3158 right = interp_exec(c, b->right, &rtype);
3159 mpq_sub(rv.num, rv.num, right.num);
3162 rv = interp_exec(c, b->left, &rvtype);
3163 right = interp_exec(c, b->right, &rtype);
3164 mpq_mul(rv.num, rv.num, right.num);
3167 rv = interp_exec(c, b->left, &rvtype);
3168 right = interp_exec(c, b->right, &rtype);
3169 mpq_div(rv.num, rv.num, right.num);
3174 left = interp_exec(c, b->left, <ype);
3175 right = interp_exec(c, b->right, &rtype);
3176 mpz_init(l); mpz_init(r); mpz_init(rem);
3177 mpz_tdiv_q(l, mpq_numref(left.num), mpq_denref(left.num));
3178 mpz_tdiv_q(r, mpq_numref(right.num), mpq_denref(right.num));
3179 mpz_tdiv_r(rem, l, r);
3180 val_init(Tnum, &rv);
3181 mpq_set_z(rv.num, rem);
3182 mpz_clear(r); mpz_clear(l); mpz_clear(rem);
3187 rv = interp_exec(c, b->right, &rvtype);
3188 mpq_neg(rv.num, rv.num);
3191 rv = interp_exec(c, b->right, &rvtype);
3192 mpq_abs(rv.num, rv.num);
3195 rv = interp_exec(c, b->right, &rvtype);
3198 left = interp_exec(c, b->left, <ype);
3199 right = interp_exec(c, b->right, &rtype);
3201 rv.str = text_join(left.str, right.str);
3204 right = interp_exec(c, b->right, &rvtype);
3208 struct text tx = right.str;
3211 if (tx.txt[0] == '-') {
3212 neg = 1; // UNTESTED
3213 tx.txt++; // UNTESTED
3214 tx.len--; // UNTESTED
3216 if (number_parse(rv.num, tail, tx) == 0)
3217 mpq_init(rv.num); // UNTESTED
3219 mpq_neg(rv.num, rv.num); // UNTESTED
3221 printf("Unsupported suffix: %.*s\n", tx.len, tx.txt); // UNTESTED
3225 ###### value functions
3227 static struct text text_join(struct text a, struct text b)
3230 rv.len = a.len + b.len;
3231 rv.txt = malloc(rv.len);
3232 memcpy(rv.txt, a.txt, a.len);
3233 memcpy(rv.txt+a.len, b.txt, b.len);
3237 ### Blocks, Statements, and Statement lists.
3239 Now that we have expressions out of the way we need to turn to
3240 statements. There are simple statements and more complex statements.
3241 Simple statements do not contain (syntactic) newlines, complex statements do.
3243 Statements often come in sequences and we have corresponding simple
3244 statement lists and complex statement lists.
3245 The former comprise only simple statements separated by semicolons.
3246 The later comprise complex statements and simple statement lists. They are
3247 separated by newlines. Thus the semicolon is only used to separate
3248 simple statements on the one line. This may be overly restrictive,
3249 but I'm not sure I ever want a complex statement to share a line with
3252 Note that a simple statement list can still use multiple lines if
3253 subsequent lines are indented, so
3255 ###### Example: wrapped simple statement list
3260 is a single simple statement list. This might allow room for
3261 confusion, so I'm not set on it yet.
3263 A simple statement list needs no extra syntax. A complex statement
3264 list has two syntactic forms. It can be enclosed in braces (much like
3265 C blocks), or it can be introduced by an indent and continue until an
3266 unindented newline (much like Python blocks). With this extra syntax
3267 it is referred to as a block.
3269 Note that a block does not have to include any newlines if it only
3270 contains simple statements. So both of:
3272 if condition: a=b; d=f
3274 if condition { a=b; print f }
3278 In either case the list is constructed from a `binode` list with
3279 `Block` as the operator. When parsing the list it is most convenient
3280 to append to the end, so a list is a list and a statement. When using
3281 the list it is more convenient to consider a list to be a statement
3282 and a list. So we need a function to re-order a list.
3283 `reorder_bilist` serves this purpose.
3285 The only stand-alone statement we introduce at this stage is `pass`
3286 which does nothing and is represented as a `NULL` pointer in a `Block`
3287 list. Other stand-alone statements will follow once the infrastructure
3298 Block -> { IN OptNL Statementlist OUT OptNL } ${ $0 = $<Sl; }$
3299 | { SimpleStatements } ${ $0 = reorder_bilist($<SS); }$
3300 | SimpleStatements ; ${ $0 = reorder_bilist($<SS); }$
3301 | SimpleStatements EOL ${ $0 = reorder_bilist($<SS); }$
3302 | IN OptNL Statementlist OUT ${ $0 = $<Sl; }$
3304 OpenBlock -> OpenScope { IN OptNL Statementlist OUT OptNL } ${ $0 = $<Sl; }$
3305 | OpenScope { SimpleStatements } ${ $0 = reorder_bilist($<SS); }$
3306 | OpenScope SimpleStatements ; ${ $0 = reorder_bilist($<SS); }$
3307 | OpenScope SimpleStatements EOL ${ $0 = reorder_bilist($<SS); }$
3308 | IN OpenScope OptNL Statementlist OUT ${ $0 = $<Sl; }$
3310 UseBlock -> { OpenScope IN OptNL Statementlist OUT OptNL } ${ $0 = $<Sl; }$
3311 | { OpenScope SimpleStatements } ${ $0 = reorder_bilist($<SS); }$
3312 | IN OpenScope OptNL Statementlist OUT ${ $0 = $<Sl; }$
3314 ColonBlock -> { IN OptNL Statementlist OUT OptNL } ${ $0 = $<Sl; }$
3315 | { SimpleStatements } ${ $0 = reorder_bilist($<SS); }$
3316 | : SimpleStatements ; ${ $0 = reorder_bilist($<SS); }$
3317 | : SimpleStatements EOL ${ $0 = reorder_bilist($<SS); }$
3318 | : IN OptNL Statementlist OUT ${ $0 = $<Sl; }$
3320 Statementlist -> ComplexStatements ${ $0 = reorder_bilist($<CS); }$
3322 ComplexStatements -> ComplexStatements ComplexStatement ${
3332 | ComplexStatement ${
3344 ComplexStatement -> SimpleStatements Newlines ${
3345 $0 = reorder_bilist($<SS);
3347 | SimpleStatements ; Newlines ${
3348 $0 = reorder_bilist($<SS);
3350 ## ComplexStatement Grammar
3353 SimpleStatements -> SimpleStatements ; SimpleStatement ${
3359 | SimpleStatement ${
3367 SimpleStatement -> pass ${ $0 = NULL; }$
3368 | ERROR ${ tok_err(c, "Syntax error in statement", &$1); }$
3369 ## SimpleStatement Grammar
3371 ###### print binode cases
3375 if (b->left == NULL) // UNTESTED
3376 printf("pass"); // UNTESTED
3378 print_exec(b->left, indent, bracket); // UNTESTED
3379 if (b->right) { // UNTESTED
3380 printf("; "); // UNTESTED
3381 print_exec(b->right, indent, bracket); // UNTESTED
3384 // block, one per line
3385 if (b->left == NULL)
3386 do_indent(indent, "pass\n");
3388 print_exec(b->left, indent, bracket);
3390 print_exec(b->right, indent, bracket);
3394 ###### propagate binode cases
3397 /* If any statement returns something other than Tnone
3398 * or Tbool then all such must return same type.
3399 * As each statement may be Tnone or something else,
3400 * we must always pass NULL (unknown) down, otherwise an incorrect
3401 * error might occur. We never return Tnone unless it is
3406 for (e = b; e; e = cast(binode, e->right)) {
3407 t = propagate_types(e->left, c, ok, NULL, rules);
3408 if ((rules & Rboolok) && t == Tbool)
3410 if (t && t != Tnone && t != Tbool) {
3414 type_err(c, "error: expected %1%r, found %2",
3415 e->left, type, rules, t);
3421 ###### interp binode cases
3423 while (rvtype == Tnone &&
3426 rv = interp_exec(c, b->left, &rvtype);
3427 b = cast(binode, b->right);
3431 ### The Print statement
3433 `print` is a simple statement that takes a comma-separated list of
3434 expressions and prints the values separated by spaces and terminated
3435 by a newline. No control of formatting is possible.
3437 `print` faces the same list-ordering issue as blocks, and uses the
3443 ##### expr precedence
3446 ###### SimpleStatement Grammar
3448 | print ExpressionList ${
3449 $0 = reorder_bilist($<2);
3451 | print ExpressionList , ${
3456 $0 = reorder_bilist($0);
3467 ExpressionList -> ExpressionList , Expression ${
3480 ###### print binode cases
3483 do_indent(indent, "print");
3487 print_exec(b->left, -1, bracket);
3491 b = cast(binode, b->right);
3497 ###### propagate binode cases
3500 /* don't care but all must be consistent */
3501 propagate_types(b->left, c, ok, NULL, Rnolabel);
3502 propagate_types(b->right, c, ok, NULL, Rnolabel);
3505 ###### interp binode cases
3511 for ( ; b; b = cast(binode, b->right))
3515 left = interp_exec(c, b->left, <ype);
3516 print_value(ltype, &left);
3517 free_value(ltype, &left);
3528 ###### Assignment statement
3530 An assignment will assign a value to a variable, providing it hasn't
3531 been declared as a constant. The analysis phase ensures that the type
3532 will be correct so the interpreter just needs to perform the
3533 calculation. There is a form of assignment which declares a new
3534 variable as well as assigning a value. If a name is assigned before
3535 it is declared, and error will be raised as the name is created as
3536 `Tlabel` and it is illegal to assign to such names.
3542 ###### declare terminals
3545 ###### SimpleStatement Grammar
3546 | Variable = Expression ${
3552 | VariableDecl = Expression ${
3560 if ($1->var->where_set == NULL) {
3562 "Variable declared with no type or value: %v",
3572 ###### print binode cases
3575 do_indent(indent, "");
3576 print_exec(b->left, indent, bracket);
3578 print_exec(b->right, indent, bracket);
3585 struct variable *v = cast(var, b->left)->var;
3586 do_indent(indent, "");
3587 print_exec(b->left, indent, bracket);
3588 if (cast(var, b->left)->var->constant) {
3589 if (v->where_decl == v->where_set) {
3591 type_print(v->type, stdout);
3596 if (v->where_decl == v->where_set) {
3598 type_print(v->type, stdout);
3605 print_exec(b->right, indent, bracket);
3612 ###### propagate binode cases
3616 /* Both must match and not be labels,
3617 * Type must support 'dup',
3618 * For Assign, left must not be constant.
3621 t = propagate_types(b->left, c, ok, NULL,
3622 Rnolabel | (b->op == Assign ? Rnoconstant : 0));
3627 if (propagate_types(b->right, c, ok, t, 0) != t)
3628 if (b->left->type == Xvar)
3629 type_err(c, "info: variable '%v' was set as %1 here.",
3630 cast(var, b->left)->var->where_set, t, rules, NULL);
3632 t = propagate_types(b->right, c, ok, NULL, Rnolabel);
3634 propagate_types(b->left, c, ok, t,
3635 (b->op == Assign ? Rnoconstant : 0));
3637 if (t && t->dup == NULL)
3638 type_err(c, "error: cannot assign value of type %1", b, t, 0, NULL);
3643 ###### interp binode cases
3646 lleft = linterp_exec(c, b->left, <ype);
3647 right = interp_exec(c, b->right, &rtype);
3649 free_value(ltype, lleft);
3650 dup_value(ltype, &right, lleft);
3657 struct variable *v = cast(var, b->left)->var;
3661 val = var_value(c, v);
3662 free_value(v->type, val);
3663 if (v->type->prepare_type)
3664 v->type->prepare_type(c, v->type, 0);
3666 right = interp_exec(c, b->right, &rtype);
3667 memcpy(val, &right, rtype->size);
3670 val_init(v->type, val);
3675 ### The `use` statement
3677 The `use` statement is the last "simple" statement. It is needed when
3678 the condition in a conditional statement is a block. `use` works much
3679 like `return` in C, but only completes the `condition`, not the whole
3685 ###### expr precedence
3688 ###### SimpleStatement Grammar
3690 $0 = new_pos(binode, $1);
3693 if ($0->right->type == Xvar) {
3694 struct var *v = cast(var, $0->right);
3695 if (v->var->type == Tnone) {
3696 /* Convert this to a label */
3699 v->var->type = Tlabel;
3700 val = global_alloc(c, Tlabel, v->var, NULL);
3706 ###### print binode cases
3709 do_indent(indent, "use ");
3710 print_exec(b->right, -1, bracket);
3715 ###### propagate binode cases
3718 /* result matches value */
3719 return propagate_types(b->right, c, ok, type, 0);
3721 ###### interp binode cases
3724 rv = interp_exec(c, b->right, &rvtype);
3727 ### The Conditional Statement
3729 This is the biggy and currently the only complex statement. This
3730 subsumes `if`, `while`, `do/while`, `switch`, and some parts of `for`.
3731 It is comprised of a number of parts, all of which are optional though
3732 set combinations apply. Each part is (usually) a key word (`then` is
3733 sometimes optional) followed by either an expression or a code block,
3734 except the `casepart` which is a "key word and an expression" followed
3735 by a code block. The code-block option is valid for all parts and,
3736 where an expression is also allowed, the code block can use the `use`
3737 statement to report a value. If the code block does not report a value
3738 the effect is similar to reporting `True`.
3740 The `else` and `case` parts, as well as `then` when combined with
3741 `if`, can contain a `use` statement which will apply to some
3742 containing conditional statement. `for` parts, `do` parts and `then`
3743 parts used with `for` can never contain a `use`, except in some
3744 subordinate conditional statement.
3746 If there is a `forpart`, it is executed first, only once.
3747 If there is a `dopart`, then it is executed repeatedly providing
3748 always that the `condpart` or `cond`, if present, does not return a non-True
3749 value. `condpart` can fail to return any value if it simply executes
3750 to completion. This is treated the same as returning `True`.
3752 If there is a `thenpart` it will be executed whenever the `condpart`
3753 or `cond` returns True (or does not return any value), but this will happen
3754 *after* `dopart` (when present).
3756 If `elsepart` is present it will be executed at most once when the
3757 condition returns `False` or some value that isn't `True` and isn't
3758 matched by any `casepart`. If there are any `casepart`s, they will be
3759 executed when the condition returns a matching value.
3761 The particular sorts of values allowed in case parts has not yet been
3762 determined in the language design, so nothing is prohibited.
3764 The various blocks in this complex statement potentially provide scope
3765 for variables as described earlier. Each such block must include the
3766 "OpenScope" nonterminal before parsing the block, and must call
3767 `var_block_close()` when closing the block.
3769 The code following "`if`", "`switch`" and "`for`" does not get its own
3770 scope, but is in a scope covering the whole statement, so names
3771 declared there cannot be redeclared elsewhere. Similarly the
3772 condition following "`while`" is in a scope the covers the body
3773 ("`do`" part) of the loop, and which does not allow conditional scope
3774 extension. Code following "`then`" (both looping and non-looping),
3775 "`else`" and "`case`" each get their own local scope.
3777 The type requirements on the code block in a `whilepart` are quite
3778 unusal. It is allowed to return a value of some identifiable type, in
3779 which case the loop aborts and an appropriate `casepart` is run, or it
3780 can return a Boolean, in which case the loop either continues to the
3781 `dopart` (on `True`) or aborts and runs the `elsepart` (on `False`).
3782 This is different both from the `ifpart` code block which is expected to
3783 return a Boolean, or the `switchpart` code block which is expected to
3784 return the same type as the casepart values. The correct analysis of
3785 the type of the `whilepart` code block is the reason for the
3786 `Rboolok` flag which is passed to `propagate_types()`.
3788 The `cond_statement` cannot fit into a `binode` so a new `exec` is
3797 struct exec *action;
3798 struct casepart *next;
3800 struct cond_statement {
3802 struct exec *forpart, *condpart, *dopart, *thenpart, *elsepart;
3803 struct casepart *casepart;
3806 ###### ast functions
3808 static void free_casepart(struct casepart *cp)
3812 free_exec(cp->value);
3813 free_exec(cp->action);
3820 static void free_cond_statement(struct cond_statement *s)
3824 free_exec(s->forpart);
3825 free_exec(s->condpart);
3826 free_exec(s->dopart);
3827 free_exec(s->thenpart);
3828 free_exec(s->elsepart);
3829 free_casepart(s->casepart);
3833 ###### free exec cases
3834 case Xcond_statement: free_cond_statement(cast(cond_statement, e)); break;
3836 ###### ComplexStatement Grammar
3837 | CondStatement ${ $0 = $<1; }$
3839 ###### expr precedence
3840 $TERM for then while do
3847 // A CondStatement must end with EOL, as does CondSuffix and
3849 // ForPart, ThenPart, SwitchPart, CasePart are non-empty and
3850 // may or may not end with EOL
3851 // WhilePart and IfPart include an appropriate Suffix
3853 // Both ForPart and Whilepart open scopes, and CondSuffix only
3854 // closes one - so in the first branch here we have another to close.
3855 CondStatement -> ForPart OptNL ThenPart OptNL WhilePart CondSuffix ${
3858 $0->thenpart = $<TP;
3859 $0->condpart = $WP.condpart; $WP.condpart = NULL;
3860 $0->dopart = $WP.dopart; $WP.dopart = NULL;
3861 var_block_close(c, CloseSequential);
3863 | ForPart OptNL WhilePart CondSuffix ${
3866 $0->condpart = $WP.condpart; $WP.condpart = NULL;
3867 $0->dopart = $WP.dopart; $WP.dopart = NULL;
3868 var_block_close(c, CloseSequential);
3870 | WhilePart CondSuffix ${
3872 $0->condpart = $WP.condpart; $WP.condpart = NULL;
3873 $0->dopart = $WP.dopart; $WP.dopart = NULL;
3875 | SwitchPart OptNL CasePart CondSuffix ${
3877 $0->condpart = $<SP;
3878 $CP->next = $0->casepart;
3879 $0->casepart = $<CP;
3881 | SwitchPart : IN OptNL CasePart CondSuffix OUT Newlines ${
3883 $0->condpart = $<SP;
3884 $CP->next = $0->casepart;
3885 $0->casepart = $<CP;
3887 | IfPart IfSuffix ${
3889 $0->condpart = $IP.condpart; $IP.condpart = NULL;
3890 $0->thenpart = $IP.thenpart; $IP.thenpart = NULL;
3891 // This is where we close an "if" statement
3892 var_block_close(c, CloseSequential);
3895 CondSuffix -> IfSuffix ${
3897 // This is where we close scope of the whole
3898 // "for" or "while" statement
3899 var_block_close(c, CloseSequential);
3901 | Newlines CasePart CondSuffix ${
3903 $CP->next = $0->casepart;
3904 $0->casepart = $<CP;
3906 | CasePart CondSuffix ${
3908 $CP->next = $0->casepart;
3909 $0->casepart = $<CP;
3912 IfSuffix -> Newlines ${ $0 = new(cond_statement); }$
3913 | Newlines ElsePart ${ $0 = $<EP; }$
3914 | ElsePart ${$0 = $<EP; }$
3916 ElsePart -> else OpenBlock Newlines ${
3917 $0 = new(cond_statement);
3918 $0->elsepart = $<OB;
3919 var_block_close(c, CloseElse);
3921 | else OpenScope CondStatement ${
3922 $0 = new(cond_statement);
3923 $0->elsepart = $<CS;
3924 var_block_close(c, CloseElse);
3928 CasePart -> case Expression OpenScope ColonBlock ${
3929 $0 = calloc(1,sizeof(struct casepart));
3932 var_block_close(c, CloseParallel);
3936 // These scopes are closed in CondSuffix
3937 ForPart -> for OpenBlock ${
3941 ThenPart -> then OpenBlock ${
3943 var_block_close(c, CloseSequential);
3947 // This scope is closed in CondSuffix
3948 WhilePart -> while UseBlock OptNL do Block ${
3952 | while OpenScope Expression ColonBlock ${
3953 $0.condpart = $<Exp;
3957 IfPart -> if UseBlock OptNL then OpenBlock ClosePara ${
3961 | if OpenScope Expression OpenScope ColonBlock ClosePara ${
3965 | if OpenScope Expression OpenScope OptNL then Block ClosePara ${
3971 // This scope is closed in CondSuffix
3972 SwitchPart -> switch OpenScope Expression ${
3975 | switch UseBlock ${
3979 ###### print exec cases
3981 case Xcond_statement:
3983 struct cond_statement *cs = cast(cond_statement, e);
3984 struct casepart *cp;
3986 do_indent(indent, "for");
3987 if (bracket) printf(" {\n"); else printf("\n");
3988 print_exec(cs->forpart, indent+1, bracket);
3991 do_indent(indent, "} then {\n");
3993 do_indent(indent, "then\n");
3994 print_exec(cs->thenpart, indent+1, bracket);
3996 if (bracket) do_indent(indent, "}\n");
4000 if (cs->condpart && cs->condpart->type == Xbinode &&
4001 cast(binode, cs->condpart)->op == Block) {
4003 do_indent(indent, "while {\n");
4005 do_indent(indent, "while\n");
4006 print_exec(cs->condpart, indent+1, bracket);
4008 do_indent(indent, "} do {\n");
4010 do_indent(indent, "do\n");
4011 print_exec(cs->dopart, indent+1, bracket);
4013 do_indent(indent, "}\n");
4015 do_indent(indent, "while ");
4016 print_exec(cs->condpart, 0, bracket);
4021 print_exec(cs->dopart, indent+1, bracket);
4023 do_indent(indent, "}\n");
4028 do_indent(indent, "switch");
4030 do_indent(indent, "if");
4031 if (cs->condpart && cs->condpart->type == Xbinode &&
4032 cast(binode, cs->condpart)->op == Block) {
4033 if (bracket) // UNTESTED
4034 printf(" {\n"); // UNTESTED
4036 printf(":\n"); // UNTESTED
4037 print_exec(cs->condpart, indent+1, bracket); // UNTESTED
4038 if (bracket) // UNTESTED
4039 do_indent(indent, "}\n"); // UNTESTED
4040 if (cs->thenpart) { // UNTESTED
4041 do_indent(indent, "then:\n"); // UNTESTED
4042 print_exec(cs->thenpart, indent+1, bracket); // UNTESTED
4046 print_exec(cs->condpart, 0, bracket);
4052 print_exec(cs->thenpart, indent+1, bracket);
4054 do_indent(indent, "}\n");
4059 for (cp = cs->casepart; cp; cp = cp->next) {
4060 do_indent(indent, "case ");
4061 print_exec(cp->value, -1, 0);
4066 print_exec(cp->action, indent+1, bracket);
4068 do_indent(indent, "}\n");
4071 do_indent(indent, "else");
4076 print_exec(cs->elsepart, indent+1, bracket);
4078 do_indent(indent, "}\n");
4083 ###### propagate exec cases
4084 case Xcond_statement:
4086 // forpart and dopart must return Tnone
4087 // thenpart must return Tnone if there is a dopart,
4088 // otherwise it is like elsepart.
4090 // be bool if there is no casepart
4091 // match casepart->values if there is a switchpart
4092 // either be bool or match casepart->value if there
4094 // elsepart and casepart->action must match the return type
4095 // expected of this statement.
4096 struct cond_statement *cs = cast(cond_statement, prog);
4097 struct casepart *cp;
4099 t = propagate_types(cs->forpart, c, ok, Tnone, 0);
4100 if (!type_compat(Tnone, t, 0))
4101 *ok = 0; // UNTESTED
4102 t = propagate_types(cs->dopart, c, ok, Tnone, 0);
4103 if (!type_compat(Tnone, t, 0))
4104 *ok = 0; // UNTESTED
4106 t = propagate_types(cs->thenpart, c, ok, Tnone, 0);
4107 if (!type_compat(Tnone, t, 0))
4108 *ok = 0; // UNTESTED
4110 if (cs->casepart == NULL)
4111 propagate_types(cs->condpart, c, ok, Tbool, 0);
4113 /* Condpart must match case values, with bool permitted */
4115 for (cp = cs->casepart;
4116 cp && !t; cp = cp->next)
4117 t = propagate_types(cp->value, c, ok, NULL, 0);
4118 if (!t && cs->condpart)
4119 t = propagate_types(cs->condpart, c, ok, NULL, Rboolok); // UNTESTED
4120 // Now we have a type (I hope) push it down
4122 for (cp = cs->casepart; cp; cp = cp->next)
4123 propagate_types(cp->value, c, ok, t, 0);
4124 propagate_types(cs->condpart, c, ok, t, Rboolok);
4127 // (if)then, else, and case parts must return expected type.
4128 if (!cs->dopart && !type)
4129 type = propagate_types(cs->thenpart, c, ok, NULL, rules);
4131 type = propagate_types(cs->elsepart, c, ok, NULL, rules);
4132 for (cp = cs->casepart;
4134 cp = cp->next) // UNTESTED
4135 type = propagate_types(cp->action, c, ok, NULL, rules); // UNTESTED
4138 propagate_types(cs->thenpart, c, ok, type, rules);
4139 propagate_types(cs->elsepart, c, ok, type, rules);
4140 for (cp = cs->casepart; cp ; cp = cp->next)
4141 propagate_types(cp->action, c, ok, type, rules);
4147 ###### interp exec cases
4148 case Xcond_statement:
4150 struct value v, cnd;
4151 struct type *vtype, *cndtype;
4152 struct casepart *cp;
4153 struct cond_statement *cs = cast(cond_statement, e);
4156 interp_exec(c, cs->forpart, NULL);
4159 cnd = interp_exec(c, cs->condpart, &cndtype);
4161 cndtype = Tnone; // UNTESTED
4162 if (!(cndtype == Tnone ||
4163 (cndtype == Tbool && cnd.bool != 0)))
4165 // cnd is Tnone or Tbool, doesn't need to be freed
4167 interp_exec(c, cs->dopart, NULL);
4170 rv = interp_exec(c, cs->thenpart, &rvtype);
4171 if (rvtype != Tnone || !cs->dopart)
4173 free_value(rvtype, &rv);
4176 } while (cs->dopart);
4178 for (cp = cs->casepart; cp; cp = cp->next) {
4179 v = interp_exec(c, cp->value, &vtype);
4180 if (value_cmp(cndtype, vtype, &v, &cnd) == 0) {
4181 free_value(vtype, &v);
4182 free_value(cndtype, &cnd);
4183 rv = interp_exec(c, cp->action, &rvtype);
4186 free_value(vtype, &v);
4188 free_value(cndtype, &cnd);
4190 rv = interp_exec(c, cs->elsepart, &rvtype);
4197 ### Top level structure
4199 All the language elements so far can be used in various places. Now
4200 it is time to clarify what those places are.
4202 At the top level of a file there will be a number of declarations.
4203 Many of the things that can be declared haven't been described yet,
4204 such as functions, procedures, imports, and probably more.
4205 For now there are two sorts of things that can appear at the top
4206 level. They are predefined constants, `struct` types, and the `main`
4207 function. While the syntax will allow the `main` function to appear
4208 multiple times, that will trigger an error if it is actually attempted.
4210 The various declarations do not return anything. They store the
4211 various declarations in the parse context.
4213 ###### Parser: grammar
4216 Ocean -> OptNL DeclarationList
4218 ## declare terminals
4225 DeclarationList -> Declaration
4226 | DeclarationList Declaration
4228 Declaration -> ERROR Newlines ${
4229 tok_err(c, // UNTESTED
4230 "error: unhandled parse error", &$1);
4236 ## top level grammar
4240 ### The `const` section
4242 As well as being defined in with the code that uses them, constants
4243 can be declared at the top level. These have full-file scope, so they
4244 are always `InScope`. The value of a top level constant can be given
4245 as an expression, and this is evaluated immediately rather than in the
4246 later interpretation stage. Once we add functions to the language, we
4247 will need rules concern which, if any, can be used to define a top
4250 Constants are defined in a section that starts with the reserved word
4251 `const` and then has a block with a list of assignment statements.
4252 For syntactic consistency, these must use the double-colon syntax to
4253 make it clear that they are constants. Type can also be given: if
4254 not, the type will be determined during analysis, as with other
4257 As the types constants are inserted at the head of a list, printing
4258 them in the same order that they were read is not straight forward.
4259 We take a quadratic approach here and count the number of constants
4260 (variables of depth 0), then count down from there, each time
4261 searching through for the Nth constant for decreasing N.
4263 ###### top level grammar
4267 DeclareConstant -> const { IN OptNL ConstList OUT OptNL } Newlines
4268 | const { SimpleConstList } Newlines
4269 | const IN OptNL ConstList OUT Newlines
4270 | const SimpleConstList Newlines
4272 ConstList -> ConstList SimpleConstLine
4274 SimpleConstList -> SimpleConstList ; Const
4277 SimpleConstLine -> SimpleConstList Newlines
4278 | ERROR Newlines ${ tok_err(c, "Syntax error in constant", &$1); }$
4281 CType -> Type ${ $0 = $<1; }$
4284 Const -> IDENTIFIER :: CType = Expression ${ {
4288 v = var_decl(c, $1.txt);
4290 struct var *var = new_pos(var, $1);
4291 v->where_decl = var;
4296 v = var_ref(c, $1.txt);
4297 tok_err(c, "error: name already declared", &$1);
4298 type_err(c, "info: this is where '%v' was first declared",
4299 v->where_decl, NULL, 0, NULL);
4303 propagate_types($5, c, &ok, $3, 0);
4308 struct value res = interp_exec(c, $5, &v->type);
4309 global_alloc(c, v->type, v, &res);
4313 ###### print const decls
4318 while (target != 0) {
4320 for (v = context.in_scope; v; v=v->in_scope)
4321 if (v->depth == 0) {
4332 struct value *val = var_value(&context, v);
4333 printf(" %.*s :: ", v->name->name.len, v->name->name.txt);
4334 type_print(v->type, stdout);
4336 if (v->type == Tstr)
4338 print_value(v->type, val);
4339 if (v->type == Tstr)
4347 ### Finally the whole `main` function.
4349 An Ocean program can currently have only one function - `main` - and
4350 that must exist. It expects an array of strings with a provided size.
4351 Following this is a `block` which is the code to execute.
4353 As this is the top level, several things are handled a bit
4355 The function is not interpreted by `interp_exec` as that isn't
4356 passed the argument list which the program requires. Similarly type
4357 analysis is a bit more interesting at this level.
4359 ###### top level grammar
4361 DeclareFunction -> MainFunction ${ {
4363 type_err(c, "\"main\" defined a second time",
4369 ###### print binode cases
4372 do_indent(indent, "func main(");
4373 for (b2 = cast(binode, b->left); b2; b2 = cast(binode, b2->right)) {
4374 struct variable *v = cast(var, b2->left)->var;
4376 print_exec(b2->left, 0, 0);
4378 type_print(v->type, stdout);
4384 print_exec(b->right, indent+1, bracket);
4386 do_indent(indent, "}\n");
4389 ###### propagate binode cases
4391 case Func: abort(); // NOTEST
4393 ###### core functions
4395 static int analyse_prog(struct exec *prog, struct parse_context *c)
4397 struct binode *bp = cast(binode, prog);
4401 struct type *argv_type;
4402 struct text argv_type_name = { " argv", 5 };
4407 argv_type = add_type(c, argv_type_name, &array_prototype);
4408 argv_type->array.member = Tstr;
4409 argv_type->array.unspec = 1;
4411 for (b = cast(binode, bp->left); b; b = cast(binode, b->right)) {
4415 propagate_types(b->left, c, &ok, argv_type, 0);
4417 default: /* invalid */ // NOTEST
4418 propagate_types(b->left, c, &ok, Tnone, 0); // NOTEST
4424 propagate_types(bp->right, c, &ok, Tnone, 0);
4429 /* Make sure everything is still consistent */
4430 propagate_types(bp->right, c, &ok, Tnone, 0);
4432 return 0; // UNTESTED
4437 static void interp_prog(struct parse_context *c, struct exec *prog,
4438 int argc, char **argv)
4440 struct binode *p = cast(binode, prog);
4448 al = cast(binode, p->left);
4450 struct var *v = cast(var, al->left);
4451 struct value *vl = var_value(c, v->var);
4461 mpq_set_ui(argcq, argc, 1);
4462 memcpy(var_value(c, t->array.vsize), &argcq, sizeof(argcq));
4463 t->prepare_type(c, t, 0);
4464 array_init(v->var->type, vl);
4465 for (i = 0; i < argc; i++) {
4466 struct value *vl2 = vl->array + i * v->var->type->array.member->size;
4469 arg.str.txt = argv[i];
4470 arg.str.len = strlen(argv[i]);
4471 free_value(Tstr, vl2);
4472 dup_value(Tstr, &arg, vl2);
4476 al = cast(binode, al->right);
4478 v = interp_exec(c, p->right, &vtype);
4479 free_value(vtype, &v);
4482 ###### interp binode cases
4484 case Func: abort(); // NOTEST
4486 ## And now to test it out.
4488 Having a language requires having a "hello world" program. I'll
4489 provide a little more than that: a program that prints "Hello world"
4490 finds the GCD of two numbers, prints the first few elements of
4491 Fibonacci, performs a binary search for a number, and a few other
4492 things which will likely grow as the languages grows.
4494 ###### File: oceani.mk
4497 @echo "===== DEMO ====="
4498 ./oceani --section "demo: hello" oceani.mdc 55 33
4504 four ::= 2 + 2 ; five ::= 10/2
4505 const pie ::= "I like Pie";
4506 cake ::= "The cake is"
4517 print "Hello World, what lovely oceans you have!"
4518 print "Are there", five, "?"
4519 print pi, pie, "but", cake
4521 A := $argv[1]; B := $argv[2]
4523 /* When a variable is defined in both branches of an 'if',
4524 * and used afterwards, the variables are merged.
4530 print "Is", A, "bigger than", B,"? ", bigger
4531 /* If a variable is not used after the 'if', no
4532 * merge happens, so types can be different
4535 double:string = "yes"
4536 print A, "is more than twice", B, "?", double
4539 print "double", B, "is", double
4544 if a > 0 and then b > 0:
4550 print "GCD of", A, "and", B,"is", a
4552 print a, "is not positive, cannot calculate GCD"
4554 print b, "is not positive, cannot calculate GCD"
4559 print "Fibonacci:", f1,f2,
4560 then togo = togo - 1
4568 /* Binary search... */
4573 mid := (lo + hi) / 2
4585 print "Yay, I found", target
4587 print "Closest I found was", mid
4592 // "middle square" PRNG. Not particularly good, but one my
4593 // Dad taught me - the first one I ever heard of.
4594 for i:=1; then i = i + 1; while i < size:
4595 n := list[i-1] * list[i-1]
4596 list[i] = (n / 100) % 10 000
4598 print "Before sort:",
4599 for i:=0; then i = i + 1; while i < size:
4603 for i := 1; then i=i+1; while i < size:
4604 for j:=i-1; then j=j-1; while j >= 0:
4605 if list[j] > list[j+1]:
4609 print " After sort:",
4610 for i:=0; then i = i + 1; while i < size:
4614 if 1 == 2 then print "yes"; else print "no"
4618 bob.alive = (bob.name == "Hello")
4619 print "bob", "is" if bob.alive else "isn't", "alive"